技术领域Technical Field
本发明涉及一种涂料的制作方法以及预测色彩数据的方法。详细而言，涉及一种基于计算机调色的涂料的制作方法以及预测涂膜的色彩数据的方法。The present invention relates to a method for preparing a coating and a method for predicting color data. Specifically, the present invention relates to a method for preparing a coating based on computer color matching and a method for predicting color data of a coating film.
另外，本发明涉及一种计算机调色系统、预测涂膜的色彩数据的系统以及用于对这些系统进行控制并使其工作的应用程序软件。In addition, the present invention relates to a computer color matching system, a system for predicting color data of a coating film, and application software for controlling and operating these systems.
背景技术Background Art
近年来，从个人喜好的多样化及提高美观的观点等考虑，作为各种工业产品，尤其作为汽车的色彩而基于金属粉或光亮性云母等光亮性颜料的光亮性色彩增加。当对这样的光亮性色彩进行修补等时，通常将含有光亮性颜料的修补用涂料涂布于修补部分。In recent years, due to the diversification of personal preferences and the improvement of aesthetics, the number of glossy colors based on glossy pigments such as metal powder and glossy mica has increased as various industrial products, especially as automobile colors. When such glossy colors are repaired, a repair paint containing a glossy pigment is usually applied to the repaired portion.
当进行光亮性色彩的修补时，不仅调整色调而且还需要对光亮感进行调整，以便修补部位等不明显。但是，当调制满足色调和光亮感这双方的光亮性色彩的修补用涂料时，即使熟练的作业者有时也需要反复经历试验和失败而试制多种修补用涂料。而且，对经验较少的作业者来讲，需要反复经历熟练的作业者以上的试验和失败而试制多种修补用涂料，成为非常困难的作业。When repairing a glossy color, it is necessary to adjust not only the color tone but also the glossiness so that the repaired part is not noticeable. However, when preparing a glossy color repair paint that satisfies both the color tone and the glossiness, even a skilled operator sometimes needs to repeatedly go through trials and failures to trial-make a variety of repair paints. Moreover, for an operator with less experience, it is very difficult to repeatedly go through trials and failures that are more than that of a skilled operator to trial-make a variety of repair paints.
由于作业者反复经历试验和失败，因此作业时间变长，在产生修补期间的长期化及效率的降低等问题的基础上，因试制出来的无法使用的修补用涂料而还产生成本及废弃的问题等。Since the operator repeatedly goes through trials and failures, the work time becomes longer, and in addition to the problems of prolonged repair period and reduced efficiency, there are also problems such as cost and disposal of the repair paint that cannot be used after trial production.
因此，为了在不受作业者的熟练度等主要原因的影响的情况下迅速且高效地制作修补用涂料，进行了各种研究。作为其中之一，进行了从通过色度计的测色而得到的作为目标的色彩的色彩数据和已知配合组成的色样本的色彩数据，使用计算机取得作为目标的色彩的配合组成的计算机色彩校正(CCM)系统的研究。但是，光亮性色彩例如含有光亮性颜料的金属涂色，具有分光反射率的角度依赖性，在现有的CCM中难以应对。Therefore, various studies have been conducted to produce repair paint quickly and efficiently without being affected by factors such as the proficiency of the operator. As one of them, a computer color calibration (CCM) system has been studied to obtain the target color combination composition using a computer from the color data of the target color obtained by color measurement with a colorimeter and the color data of a color sample with a known combination composition. However, bright colors such as metallic paints containing bright pigments have an angle dependence of spectral reflectance, which is difficult to cope with in existing CCMs.
另外，在目前为止的修补用涂料的调色中，经常会发生在得到近似的涂料调色配合(有时会简化为近似配合)之后，以此为出发点需要进一步进行微调色，尤其在成为光亮性色彩的修补用涂料的调制中，通过现有的CCM会在最近似配合的精度、再现性上存在极限，依然会经常发生作业者反复经历调色的试验和失败的工时，在迅速且高效地制作修补用涂料上存在问题。Furthermore, in the color matching of repair paints to date, it often happens that after obtaining an approximate paint color matching (sometimes simplified to an approximate matching), further fine-tuning of the color is required based on this as a starting point. In particular, in the preparation of repair paints to become bright colors, there are limits to the accuracy and reproducibility of the most approximate matching through existing CCMs, and operators still often go through repeated color matching experiments and failed work, which poses problems in quickly and efficiently producing repair paints.
专利文献1中公开有如下方法，在利用计算机的配色系统中，通过分光光度计决定未知的色板的全光谱反射率，将该反射率数据发往计算机，计算机对表示颜料的K值(光吸收系数)及S值(光散射系数)的预先记录的数据进行数学处理，进行理论性配色。Patent document 1 discloses the following method, in which, in a color matching system using a computer, the full spectral reflectance of an unknown color plate is determined by a spectrophotometer, and the reflectance data is sent to a computer. The computer mathematically processes the pre-recorded data representing the K value (light absorption coefficient) and S value (light scattering coefficient) of the pigment to perform theoretical color matching.
专利文献2中公开有由色度计、微观光亮感测定仪、计算机所构成的计算机调色装置以及使用其的调色方法，该计算机中登记有多种涂料配合、对应于该各涂料配合的色数据、微观光亮感数据以及多种原色涂料的色特性数据、微观光亮感特性数据且配色计算逻辑发挥作用。Patent document 2 discloses a computer color matching device composed of a colorimeter, a microscopic brightness measuring instrument, and a computer, and a color matching method using the same, wherein the computer registers a plurality of paint combinations, color data corresponding to each paint combination, microscopic brightness data, and color characteristic data of a plurality of primary color paints, and microscopic brightness characteristic data, and the color matching calculation logic functions.
专利文献3中公开有由色度计、微观光亮感样本色票、计算机所构成的调色方法，该计算机中登记有多种涂料配合、对应于该各涂料配合的色数据、微观光亮感数据以及多种原色涂料的色特性数据、微观光亮感特性数据且配色计算逻辑发挥作用。Patent document 3 discloses a color matching method consisting of a colorimeter, microscopic brightness sample color chips, and a computer, in which a plurality of paint combinations, color data corresponding to each paint combination, microscopic brightness data, and color characteristic data of a plurality of primary color paints, microscopic brightness characteristic data, are registered, and color matching calculation logic is in effect.
专利文献4中公开有如下方法，作为修补用涂料的最终微调色而有用的金属系涂色的调色方法，根据目视时的正面颜色与背面颜色的不同点，当改变金属涂料的调色配合时，使用基于涂色的正面颜色与背面颜色比例的特征以及正面颜色的亮度不发生变化的各金属原色涂料的配合变换指数。Patent document 4 discloses a method for coloring metallic paint useful as a final fine-tuning of a repair paint, which uses a matching transformation index of each metallic primary color paint based on the characteristics of the ratio of the front color and the back color of the paint and the brightness of the front color without changing when changing the color matching of the metallic paint according to the difference between the front color and the back color when viewed visually.
专利文献5中公开有如下方法，通过基于在涂料分配组成中使用的色成分的人工神经网络，决定或预测涂料的视觉上的质感参数。但是，关于分光反射特性等的光学特性，通过对已知的涂料分配组成的物理模型进行决定或预测。当对难以预测光学特性的光亮性色彩调制修补用涂料时，可预测到会需要较多的试验和失败。Patent document 5 discloses a method for determining or predicting visual texture parameters of a coating by using an artificial neural network based on color components used in the coating composition. However, optical properties such as spectral reflection properties are determined or predicted by using a physical model of a known coating composition. When modulating a repair coating for a glossy color whose optical properties are difficult to predict, it is expected that more trials and failures will be required.
专利文献6中记载有如下涂色的检索方法，在光亮性涂膜等中，为了决定具有所希望的质感的涂色或属于所希望的色类别的涂色，采用了涂色数据库制作方法及该数据库，其中包括使用涂色的分光反射率或微观光亮感等数据而使神经网络学习的步骤。但是，关于当对难以预测光学特性的光亮性色彩调制修补用涂料时的具体操作等，并未进行任何公开。Patent Document 6 discloses a method for searching for paints, which uses a method for creating a paint database and the database to determine a paint having a desired texture or a paint belonging to a desired color category in a glossy coating film, and includes a step of learning a neural network using data such as spectral reflectance or microscopic brightness of the paint. However, no specific operation is disclosed when modulating a repair paint for a glossy color whose optical properties are difficult to predict.
专利文献Patent Literature
专利文献1：日本国特公昭50-28190号公报Patent Document 1: Japanese Patent Publication No. 50-28190
专利文献2：日本国特开2001-221690号公报Patent Document 2: Japanese Patent Application Publication No. 2001-221690
专利文献3：国际公开第2002/004567号公报Patent Document 3: International Publication No. 2002/004567
专利文献4：日本国特开2004-224966号公报Patent Document 4: Japanese Patent Application Publication No. 2004-224966
专利文献5：日本国特表2019-500588号公报Patent Document 5: Japanese Patent Application No. 2019-500588
专利文献6：国际公开第2008/156147号公报Patent Document 6: International Publication No. 2008/156147
发明内容Summary of the invention
本发明的第1目的为提供一种涂料的制作方法，其用于得到包含光学特性难以预测的光亮性色彩的多种多样色彩的涂色，其不受作业者的熟练度的影响，而是基于通过较少的试制次数就可以结束调色的计算机调色。The first object of the present invention is to provide a method for preparing a coating, which is used to obtain a variety of colors including bright colors whose optical properties are difficult to predict. The method is not affected by the operator's proficiency, but is based on computer color matching that can complete the color matching through a small number of trial productions.
本发明的第2目的为提供一种预测涂膜的色彩数据的方法，关于含有光亮性颜料等的多种多样组成的涂料的涂膜，能够高精度地预测其色彩数据。A second object of the present invention is to provide a method for predicting color data of a coating film, which can accurately predict the color data of a coating film of a coating material having a variety of compositions including bright pigments and the like.
本发明的第3目的为提供一种计算机调色系统，其用于调制用于取得包含光学特性难以预测的光亮性色彩的多种多样色彩的涂色的涂料，其不受作业者的熟练度的影响，而是通过较少的试制次数就可以结束调色。The third object of the present invention is to provide a computer color matching system for modulating paint for obtaining a variety of colors including bright colors whose optical properties are difficult to predict, which is not affected by the operator's proficiency and can complete the color matching with a smaller number of trial productions.
本发明的第4目的为提供一种预测涂膜的色彩数据的系统，关于含有光亮性颜料等的多种多样组成的涂料的涂膜，能够高精度地预测其色彩数据。A fourth object of the present invention is to provide a system for predicting color data of a coating film, which can predict the color data of a coating film of a coating material having a variety of compositions including bright pigments, etc. with high accuracy.
本发明者为了解决上述课题进行了潜心研究，发现了通过做成以下的构成能够解决上述课题，以至于完成了本发明。The present inventors have conducted intensive studies to solve the above-mentioned problems, and have found that the above-mentioned problems can be solved by the following configuration, thereby completing the present invention.
具体而言，如同以下所述。Specifically, it is as follows.
项1：一种涂料的制作方法，基于使用具备数据库和计算机的装置的计算机调色，Item 1: A method for producing a coating, based on computer color matching using a device having a database and a computer,
在该数据库中登记有1种以上的组合物C1～Cn(n为2以上的整数)的各自配合组成数据Y1～Yn及分别对应于各配合组成数据的色彩数据X1～Xn，In the database, there are registered respective blending composition data Y1 to Yn of one or more compositions C1 to Cn (n is an integer greater than or equal to 2) and color data X1 to Xn corresponding to each blending composition data.
在该计算机中利用登记在该数据库中的数据的配色计算逻辑发挥作用，The computer uses the color matching calculation logic of the data registered in the database to perform the calculation.
包括下述S101～S111工序。The process includes the following steps S101 to S111.
S101是使用登记在所述数据库中的数据，向所述计算机输入学习用数据的工序，S101 is a step of inputting learning data into the computer using the data registered in the database.
S102是使用所述学习用数据进行机器学习，生成包含从色彩数据X推断配合组成数据Y的人工智能模型的至少1种的已学习的人工智能模型的工序，S102 is a step of performing machine learning using the learning data to generate at least one learned artificial intelligence model including an artificial intelligence model that infers the combination composition data Y from the color data X.
S103是取得配合组成Yp为未知的作为目标的色彩的色彩数据(Xp)的工序，S103 is a step of obtaining color data (Xp) of a target color whose combination composition Yp is unknown.
S104是向所述计算机输入所述色彩数据Xp的工序，S104 is a step of inputting the color data Xp into the computer.
S105是使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，将从色彩数据Xp预测的预测配合组成数据Ya1，作为将所述组合物C1～Cn的1种以上作为成分的配合组成数据而取得的工序，S105 is a step of obtaining predicted combination composition data Ya1 predicted from color data Xp as combination composition data having one or more of the compositions C1 to Cn as components using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model.
S106是使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，取得从预测配合组成数据Ya1预测的预测色彩数据Xa1，同时与所述色彩数据Xp进行比较而判定合格与否的工序，S106 is a process of obtaining predicted color data Xa1 predicted from predicted combination composition data Ya1 by using the learned artificial intelligence model and/or prediction formula other than the artificial intelligence model, and comparing it with the color data Xp to determine whether it is qualified or not.
S107是当在所述S106工序中不合格时，使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，将从色彩数据Xp预测的不同于到此为止的预测配合组成数据的预测配合组成数据Yai，作为将所述组合物C1～Cn的1种以上作为成分的配合组成数据而取得，之后使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，取得从预测配合组成数据Yai预测的预测色彩数据Xai，同时达到合格为止反复进行通过与所述色彩数据Xp的比较而判定合格与否的工序，S107 is a process of, when the process S106 fails, using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, obtaining predicted combination composition data Yai predicted from the color data Xp, which is different from the predicted combination composition data so far, as combination composition data having one or more of the compositions C1 to Cn as components, and then using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, obtaining predicted color data Xai predicted from the predicted combination composition data Yai, and repeatedly performing the process of judging whether the result is qualified by comparing with the color data Xp until the result is qualified,
S108是当在所述S106或S107工序的任意一个中合格时，取得合格配合组成数据Yap1的工序，S108 is a step of obtaining qualified combination composition data Yap1 when either S106 or S107 is qualified.
S109是根据所述合格配合组成数据Yap1调制实际候补涂料CMap1，得到该实际候补涂料CMap1的涂装板而取得实测色彩数据Xap1的工序，S109 is a step of modulating the actual candidate paint CMap1 based on the qualified combination composition data Yap1, obtaining a painted plate of the actual candidate paint CMap1, and obtaining the measured color data Xap1.
S110是通过所述色彩数据Xp与所述实测色彩数据Xap1的比较及/或作为所述目标的色彩与所述实际候补涂料CMap1的涂装板的色彩的比较，判定合格与否的工序，S110 is a step of judging whether the color data Xp is acceptable or not by comparing the color data Xp with the measured color data Xap1 and/or comparing the target color with the color of the coated plate of the actual candidate paint CMap1.
S111是当在所述S110工序中不合格时，达到合格为止将所述S105～S110工序或S107～S110工序反复进行的工序。S111 is a step of repeating the steps S105 to S110 or the steps S107 to S110 until the product becomes acceptable if the product fails in the step S110.
项2：一种涂料的制作方法，基于使用具备数据库和计算机的装置的计算机调色，在该数据库中至少登记有1种以上的组合物的配合组成数据Y及对应的色彩数据X，在该计算机中利用登记在该数据库中的数据的配色计算逻辑发挥作用，包括下述S201～S211工序。Item 2: A method for preparing a coating, based on computer color matching using a device equipped with a database and a computer, wherein at least one combination composition data Y and corresponding color data X of the composition are registered in the database, and a color matching calculation logic of the data registered in the database is used in the computer, comprising the following steps S201 to S211.
S201是使用登记在所述数据库中的数据，向所述计算机输入学习用数据的工序，S201 is a step of inputting learning data into the computer using the data registered in the database.
S202是使所述学习用数据进行机器学习，生成包含从配合组成数据Y推断色彩数据X的人工智能模型的至少1种的已学习的人工智能模型的工序，S202 is a step of subjecting the learning data to machine learning to generate at least one learned artificial intelligence model including an artificial intelligence model for inferring color data X from combination composition data Y.
S203是取得作为目标的色彩的目标色彩数据Xt的工序，S203 is a step of obtaining target color dataXt of a target color.
S204是向所述计算机输入所述目标色彩数据Xt的工序，S204 is a step of inputting the target color dataXt into the computer.
S205是通过使用计算机的检索，取得近似于所述目标色彩数据Xt的检索色彩数据Xn1及对应于检索色彩数据Xn1的近似配合组成数据Yn1，同时对所述目标色彩数据Xt与所述检索色彩数据Xn1进行比较而判定合格与否的工序，S205 is a process of obtaining search color dataXn1 similar to the target color dataXt and approximate combination composition dataYn1 corresponding to the search color dataXn1 by searching with a computer, and comparing the target color dataXt with the search color dataXn1 to determine whether the color data is acceptable or not.
S206是当在所述S205工序中不合格时，使用计算机取得预测为提供目标色彩数据Xt的候补配合组成数据Yni，之后使用所述至少1种已学习的人工智能模型及/或所述人工智能模型以外的预测式，取得从候补配合组成数据Yni预测的预测色彩数据Xni，同时对所述色彩数据Xt与所述预测色彩数据Xni进行比较而判定合格与否的工序，S206 is a step of obtaining candidate combination composition data Yni predicted to provide target color data Xt using a computer when the color data fails in the step S205, and then obtaining predicted color data Xni predicted from the candidate combination composition data Yni using the at least one learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, and comparing the color data Xt with the predicted color data Xni to determine whether the color data fails.
S207是当在所述S206工序中不合格时，使用计算机取得预测为提供目标色彩数据Xt的候补配合组成数据Yni，之后使用所述至少1种已学习的人工智能模型及/或所述人工智能模型以外的预测式，取得从候补配合组成数据Yni预测的预测色彩数据Xni，同时对所述色彩数据Xt与所述预测色彩数据Xni进行比较，达到合格为止将判定合格与否的工序反复进行的工序，S207 is a process in which, when the process S206 fails, candidate combination composition data Yni predicted to provide target color data Xt are obtained using a computer, and then predicted color data Xni predicted from the candidate combination composition data Yni are obtained using the at least one learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, and the color data Xt is compared with the predicted color data Xni , and the process of judging whether the color data X t is qualified is repeated until the color data X t is qualified.
S208是当在所述S205～S207工序的任意一个中合格时，取得合格配合组成数据YC1的工序，S208 is a step of obtaining qualified combination composition data YC1 when any one of the steps S205 to S207 is qualified.
S209是根据所述合格配合组成数据YC1调制实际候补涂料CMCi，得到该实际候补涂料CMCi的涂装板而取得实测色彩数据XCi的工序，S209 is a step of modulating the actual candidate paint CMCi according to the qualified combination composition data YC1 , obtaining a painted plate of the actual candidate paint CMCi and acquiring the measured color data XCi .
S210是通过所述色彩数据Xt与所述实测色彩数据XCi的比较及/或作为所述目标的色彩与所述实际候补涂料CMCi的涂装板的色彩的比较，判定合格与否的工序，S210 is a step of judging whether the color dataXt is acceptable or not by comparing the color data Xt with the measured color dataXci and/or comparing the target color with the color of the plate coated with the actual candidate paintCMci .
S211是当在所述S210工序中不合格时，将所述S206～S210工序反复进行的工序。S211 is a step of repeating the steps S206 to S210 when the product fails in the step S210.
项3：一种预测涂膜的色彩数据的方法，使用具备数据库和计算机的装置，在该数据库中至少登记有1种以上的组合物的配合组成数据Y及对应的色彩数据X，在该计算机中利用登记在该数据库中的数据的配色计算逻辑发挥作用，所述方法包括下述S301～S309工序。Item 3: A method for predicting the color data of a coating film, using a device having a database and a computer, wherein at least one composition data Y and corresponding color data X of the composition are registered in the database, and the color matching calculation logic of the data registered in the database is used in the computer, and the method includes the following steps S301 to S309.
S301是使用登记在所述数据库中的数据，向所述计算机输入学习用数据的工序，S301 is a step of inputting learning data into the computer using the data registered in the database.
S302是使所述学习用数据进行机器学习，生成包含从配合组成数据Y推断色彩数据X的人工智能模型的至少1种的已学习的人工智能模型的工序，S302 is a step of subjecting the learning data to machine learning to generate at least one learned artificial intelligence model including an artificial intelligence model for inferring color data X from combination composition data Y.
S303是取得预测涂膜的色彩数据的涂料CMt的配合组成数据YCM的工序，S304是向所述计算机输入所述配合组成数据YCM的工序，S303 is a step of obtaining the mixing composition data YCM of the paint CMt for predicting the color data of the coating film, and S304 is a step of inputting the mixing composition data YCM into the computer.
S305是根据需要，通过使用计算机的检索，取得对应于所述配合组成数据YCM的检索色彩数据Xn1的工序，S305 is a step of obtaining the search color dataXn1 corresponding to the combination composition dataYCM by searching with a computer as needed.
S306是当在所述S305工序中未检索到对应的检索色彩数据Xn1时，或者当并未进行所述S305工序时，使用所述至少1种已学习的人工智能模型或所述至少1种已学习的人工智能模型和所述人工智能模型以外的预测式，从所述配合组成数据YCM取得预测色彩数据Xm1的工序，S306 is a step of obtaining predicted color dataXm1 from the coordination composition data YCM using the at least one learned artificial intelligence model or the at least one learned artificial intelligence model and a prediction formula other than the artificial intelligence model when the corresponding search color dataXn1 is not found in theS305 step or when the S305 step is not performed.
S307是根据需要，取得涂装有所述涂料CMt的涂装板的实测色彩数据XCM，与所述预测色彩数据Xm1进行比较的工序。S307 is a step of acquiring the measured color data XCM of the painted plate painted with the paint CMt as needed, and comparing the data with the predicted color data Xm1 .
项4：根据项1所记载的涂料的制作方法，所述S105工序及/或S107工序包括，使用多标签分类将从色彩数据Xp预测的预测配合组成数据Ya1及/或Yai，作为将所述组合物C1～Cn的1种以上作为成分的配合组成数据而取得的工序。Item 4: According to the method for preparing the coating described in Item 1, the S105 process and/or the S107 process includes a process of obtaining the predicted combination composition data Ya1 and/or Yai predicted from the color data Xp as combination composition data with one or more of the compositions C1 to Cn as components using multi-label classification.
项5：根据项1或4所记载的涂料的制作方法，在所述S105工序中取得的预测配合组成数据Ya1及/或在所述S107工序中取得的预测配合组成数据Yai为将所述组合物C1～Cn的15种以下作为成分的配合组成数据，而且含有金属颜料的组合物为5种以下，含有珠光颜料的组合物为5种以下。Item 5: According to the coating preparation method described in Item 1 or 4, the predicted combination composition data Ya1 obtained in the S105 process and/or the predicted combination composition data Yai obtained in the S107 process are combination composition data with 15 or less of the compositions C1 to Cn as components, and the number of compositions containing metallic pigments is 5 or less, and the number of compositions containing pearlescent pigments is 5 or less.
项6：根据项2所记载的涂料的制作方法，当在所述S211工序中不合格时，将所述预测色彩数据Xni与所述实测色彩数据XCi的差分Δ作为修正系数α而输入到计算机，之后反复进行所述S206～S211工序。Item 6: The coating production method according to Item 2, wherein when the S211 step fails, the difference Δ between the predicted color dataXni and the measured color dataXci is input into a computer as a correction coefficient α, and then the S206 to S211 steps are repeated.
项7：根据项1、2、4～6的任意一项所述的涂料的制作方法或项3所述的预测涂膜的色彩数据的方法，登记在所述数据库中的1种以上的组合物的配合组成数据Y及对应的色彩数据X包含实测数据或者包含实测数据和根据实测数据算出的数据。Item 7: According to the method for preparing a coating as described in any one of Items 1, 2, 4 to 6 or the method for predicting color data of a coating film as described in Item 3, the compounding composition data Y of one or more compositions registered in the database and the corresponding color data X include measured data or include measured data and data calculated based on the measured data.
项8：根据项1、2、4～7的任意一项所记载的涂料的制作方法，Item 8: A method for producing a coating according to any one of Items 1, 2, 4 to 7,
在所述S102工序或所述S202工序中，生成已学习的人工智能模型的工序包括：In the step S102 or the step S202, the step of generating the learned artificial intelligence model includes:
(i)作为学习用数据而使用并不含有光亮性颜料的组合物的1种以上所涉及的各配合组成数据Y及各色彩数据X，使人工智能模型学习的工序；(i) a step of using, as learning data, each blending composition data Y and each color data X related to one or more compositions that do not contain a bright pigment to cause an artificial intelligence model to learn;
及(ii)作为学习用数据而使用含有光亮性颜料的组合物的1种以上所涉及的各配合组成数据Y及各色彩数据X，使人工智能模型学习的工序。and (ii) a step of causing an artificial intelligence model to learn using, as learning data, each blending composition data Y and each color data X concerning one or more compositions containing a bright pigment.
项9：根据项1、2、4～8的任意一项所记载的涂料的制作方法，Item 9: A method for producing a coating according to any one of Items 1, 2, 4 to 8,
在所述S102工序或所述S202工序中，生成已学习的人工智能模型的工序包括，In the step S102 or the step S202, the step of generating the learned artificial intelligence model includes:
作为学习用数据而使用选自组合物中的光反射性颜料的含量、光干涉性颜料的含量、定向控制剂的含量、组合物中的光反射性颜料的各色相的含量、光干涉性颜料的各色相的含量、着色剂的各色相的含量及这些含量的2个以上总计的1种以上的数据及/或包含于组合物的色料的形状数据，As learning data, one or more data selected from the group consisting of the content of the light reflective pigment in the composition, the content of the light interference pigment, the content of the orientation control agent, the content of each color phase of the light reflective pigment in the composition, the content of each color phase of the light interference pigment, the content of each color phase of the colorant, and the total of two or more of these contents and/or shape data of the colorant contained in the composition are used,
使人工智能模型学习的工序。The process of making artificial intelligence models learn.
项10：根据项1、2、4～9的任意一项所记载的涂料的制作方法，在所述S103工序中的色彩数据Xp或在所述S203工序中的目标色彩数据Xt为含有光亮性颜料的涂膜的色彩数据。Item 10: The coating production method according to any one of Items 1, 2, 4 to 9, wherein the color data Xp in the step S103 or the target color dataXt in the step S203 is color data of a coating film containing a bright pigment.
项11：根据项1、2、4～10的任意一项所记载的涂料的制作方法，当在进行所述判定的工序中不合格时，以使用人工智能模型以外的预测式的方式进行切换的工序包括在所述反复的工序中。Item 11: The coating production method according to any one of Items 1, 2, and 4 to 10, wherein when the determination step fails, the repeated steps include a step of switching to a method using a prediction formula other than an artificial intelligence model.
项12：根据项1、2、4～11的任意一项所记载的涂料的制作方法，在进行所述判定的工序中，判定使用计算机。Item 12: The method for producing a coating material according to any one of Items 1, 2, and 4 to 11, wherein in the step of performing the determination, the determination is performed using a computer.
项13：根据项1、2、4～12的任意一项所记载的涂料的制作方法或项3所述的预测涂膜的色彩数据的方法，在车辆的修补涂装中被使用。Item 13: The method for producing a coating material according to any one of Items 1, 2, 4 to 12 or the method for predicting color data of a coating film according to Item 3, which is used in vehicle repair painting.
项14：一种计算机调色系统，具备：Item 14: A computer color grading system comprising:
数据库，登记有1种以上的组合物C1～Cn(n为2以上的整数)的各自配合组成数据Y1～Yn和分别对应于各配合组成数据的色彩数据X1～Xn；A database registering respective combination composition data Y1 to Yn of one or more combinations C1 to Cn (n is an integer greater than or equal to 2) and color data X1 to Xn corresponding to each combination composition data;
及计算机，利用登记在该数据库中的数据的配色计算逻辑发挥作用，and computers, using the color matching calculation logic of the data registered in the database,
所述系统包括下述手段S401～S411。The system includes the following means S401 to S411.
S401是使用登记在所述数据库中的数据，向所述计算机输入学习用数据的手段，S401 is a means for inputting learning data into the computer using the data registered in the database.
S402是使所述学习用数据进行机器学习，生成包含从色彩数据X推断配合组成数据Y的人工智能模型的至少1种的已学习的人工智能模型的手段，S402 is a means for subjecting the learning data to machine learning to generate at least one learned artificial intelligence model including an artificial intelligence model for inferring the combination composition data Y from the color data X,
S403是取得配合组成Yp为未知的作为目标的色彩的色彩数据Xp的手段，S403 is a means for obtaining color data Xp of a target color whose matching composition Yp is unknown.
S404是向所述计算机输入所述色彩数据Xp的手段，S404 is a means of inputting the color data Xp into the computer.
S405是使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，将从色彩数据Xp预测的预测配合组成数据Ya1，作为将所述组合物C1～Cn的1种以上作为成分的配合组成数据而取得的手段，S405 is a means for obtaining the predicted combination composition data Ya1 predicted from the color data Xp as combination composition data having one or more of the compositions C1 to Cn as components, using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model.
S406是使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，取得从预测配合组成数据Ya1预测的预测色彩数据Xa1，同时与所述色彩数据Xp进行比较而判定合格与否的手段，S406 is a means for obtaining predicted color data Xa1 predicted from predicted combination composition data Ya1 by using the learned artificial intelligence model and/or prediction formula other than the artificial intelligence model, and comparing it with the color data Xp to determine whether it is qualified or not.
S407是当在所述手段S406中不合格时，使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，将从色彩数据Xp预测的不同于到此为止的预测配合组成数据的预测配合组成数据Yai，作为将所述组合物C1～Cn的1种以上作为成分的配合组成数据而取得，之后使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，取得从预测配合组成数据Yai预测的预测色彩数据Xai，同时达到合格为止反复进行通过与所述色彩数据Xp的比较而判定合格与否的手段，S407 is a means for repeatedly determining whether the result is qualified by comparing with the color data Xp until the result is qualified, using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model to obtain predicted combination composition data Yai predicted from the color data Xp as combination composition data having one or more of the compositions C1 to Cn as components, and then using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model to obtain predicted color data Xai predicted from the predicted combination composition data Yai, and repeatedly performing the comparison with the color data Xp to determine whether the result is qualified until the result is qualified,
S408是当在所述手段S406或S407的任意一个中合格时，取得合格配合组成数据Yap1的手段，S408 is a means for obtaining qualified combination composition data Yap1 when any one of the means S406 or S407 is qualified.
S409是根据所述合格配合组成数据Yap1调制实际候补涂料CMap1，得到该实际候补涂料CMap1的涂装板而取得实测色彩数据Xap1的手段，S409 is a means for modulating the actual candidate paint CMap1 according to the qualified combination composition data Yap1, obtaining a coating plate of the actual candidate paint CMap1 and obtaining the measured color data Xap1.
S410是通过所述色彩数据Xp与所述实测色彩数据Xap1的比较及/或作为所述目标的色彩与所述实际候补涂料CMap1的涂装板的色彩的比较，判定合格与否的手段，S410 is a means for judging pass/fail by comparing the color data Xp with the measured color data Xap1 and/or comparing the target color with the color of the painted plate of the actual candidate paint CMap1.
S411是当在所述手段S410中不合格时，达到合格为止将所述手段S405～S410或S407～S410反复进行的手段。S411 is a means of repeatedly performing the above-mentioned means S405 to S410 or S407 to S410 until the above-mentioned means S410 fails to pass.
项15：一种计算机调色系统，具备：数据库，至少登记有1种以上的组合物的配合组成数据Y及对应的色彩数据X；及计算机，利用登记在该数据库中的数据的配色计算逻辑发挥作用，所述系统包括下述手段S501～S511。Item 15: A computer color matching system comprising: a database in which at least one combination of composition data Y and corresponding color data X are registered; and a computer that operates using the color matching calculation logic of the data registered in the database, the system including the following means S501 to S511.
S501是使用登记在所述数据库中的数据，向所述计算机输入学习用数据的手段，S501 is a means for inputting learning data into the computer using the data registered in the database.
S502是使所述学习用数据进行机器学习，生成包含从配合组成数据Y推断色彩数据X的人工智能模型的至少1种的已学习的人工智能模型的手段，S502 is a means for subjecting the learning data to machine learning to generate at least one learned artificial intelligence model including an artificial intelligence model for inferring color data X from combination composition data Y,
S503是取得作为目标的色彩的目标色彩数据Xt的手段，S503 is a means of obtaining target color dataXt of the target color.
S504是向所述计算机输入所述目标色彩数据Xt的手段，S504 is a means of inputting the target color dataXt into the computer.
S505是通过使用计算机的检索，取得近似于所述目标色彩数据Xt的检索色彩数据Xn1及对应于检索色彩数据Xn1的近似配合组成数据Yn1，同时对所述目标色彩数据Xt与所述检索色彩数据Xn1进行比较而判定合格与否的手段，S505 is a means of obtaining search color dataXn1 similar to the target color dataXt and approximate matching composition dataYn1 corresponding to the search color dataXn1 by searching with a computer, and comparing the target color dataXt with the search color dataXn1 to determine whether the target color data Xt is qualified or not.
S506是当在所述手段S505中不合格时，使用计算机取得预测为提供目标色彩数据Xt的候补配合组成数据Yni，之后使用所述至少1种已学习的人工智能模型及/或所述人工智能模型以外的预测式，取得从候补配合组成数据Yni预测的预测色彩数据Xni，同时对所述色彩数据Xt与所述预测色彩数据Xni进行比较而判定合格与否的手段，S506 is a means for obtaining candidate combination composition data Yni predicted to provide target color data Xt by using a computer when the method S505 fails, and then obtaining predicted color data Xni predicted from the candidate combination composition data Yni by using the at least one learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, and comparing the color data Xt with the predicted color data Xni to determine whether the color data X t fails.
S507是当在所述手段S506中不合格时，使用计算机取得预测为提供目标色彩数据Xt的候补配合组成数据Yni，之后使用所述至少1种已学习的人工智能模型及/或所述人工智能模型以外的预测式，取得从候补配合组成数据Yni预测的预测色彩数据Xni，同时对所述色彩数据Xt与所述预测色彩数据Xni进行比较，达到合格为止将判定合格与否的手段反复进行的手段，S507 is a means for obtaining candidate combination composition data Yni predicted to provide target color data Xt by using a computer when the method S506 fails, and then obtaining predicted color data Xni predicted from the candidate combination composition data Yni by using the at least one learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, and comparing the color data Xt with the predicted color data Xni , and repeating the means of judging whether the color data X t is qualified until the color data X t is qualified.
S508是当在所述手段S505～S507的任意一个中合格时，取得合格配合组成数据YC1的手段，S508 is a means for obtaining qualified combination composition data YC1 when any of the above-mentioned means S505 to S507 is qualified.
S509是根据所述合格配合组成数据YC1调制实际候补涂料CMCi，得到该实际候补涂料CMCi的涂装板而取得实测色彩数据XCi的手段，S509 is a means for modulating the actual candidate paint CMCi according to the qualified combination composition data YC1 to obtain a coating plate of the actual candidate paint CMCi and obtain the measured color data XCi .
S510是通过所述色彩数据Xt与所述实测色彩数据XCi的比较及/或作为所述目标的色彩与所述实际候补涂料CMCi的涂装板的色彩的比较，判定合格与否的手段，S510 is a means for judging pass/fail by comparing the color dataXt with the measured color dataXCi and/or comparing the target color with the color of the plate coated with the actual candidate paintCMCi .
S511是当在所述手段S510中不合格时，将所述手段S506～S510反复进行的手段。S511 is a means of repeatedly performing the means S506 to S510 when the means S510 fails.
项16：一种预测涂膜的色彩数据的系统，具备：数据库，至少登记有1种以上的组合物的配合组成数据Y及对应的色彩数据X；及计算机，利用登记在该数据库中的数据的配色计算逻辑发挥作用，所述系统包括下述手段S601～S609。Item 16: A system for predicting the color data of a coating film, comprising: a database in which at least one composition data Y and corresponding color data X are registered; and a computer that operates using the color matching calculation logic of the data registered in the database, the system including the following means S601 to S609.
S601是使用登记在所述数据库中的数据，向所述计算机输入学习用数据的手段，S601 is a means for inputting learning data into the computer using the data registered in the database.
S602是使所述学习用数据进行机器学习，生成包含从配合组成数据Y推断色彩数据X的人工智能模型的至少1种的已学习的人工智能模型的手段，S602 is a means for subjecting the learning data to machine learning to generate at least one learned artificial intelligence model including an artificial intelligence model for inferring color data X from combination composition data Y,
S603是取得预测涂膜的色彩数据的涂料CMt的配合组成数据YCM的手段，S604是向所述计算机输入所述配合组成数据YCM的手段，S603 is a means for obtaining the mixing composition data YCM of the paint CMt for predicting the color data of the coating film, and S604 is a means for inputting the mixing composition data YCM into the computer.
S605是根据需要，通过使用计算机的检索，取得对应于所述配合组成数据YCM的检索色彩数据Xn1的手段，S605 is a means for obtaining the search color dataXn1 corresponding to the coordination composition dataYCM by searching using a computer as needed.
S606是当在所述手段S605中未检索到对应的检索色彩数据Xn1时，或者当并未进行所述手段S605时，使用所述至少1种已学习的人工智能模型或所述至少1种已学习的人工智能模型和所述人工智能模型以外的预测式，从所述配合组成数据YCM取得预测色彩数据Xm1的手段，S606 is a means for obtaining predicted color dataXm1 from the coordination composition data YCM using the at least one learned artificial intelligence model or the at least one learned artificial intelligence model and a prediction formula other than the artificial intelligence model when the corresponding search color dataXn1 is not found in the meansS605 or when the means S605 is not performed.
S607是根据需要，取得涂装有所述涂料CMt的涂装板的实测色彩数据XCM，与所述预测色彩数据Xm1进行比较的手段。S607 is a means for obtaining the measured color data XCM of the painted plate painted with the paint CMt as needed, and comparing it with the predicted color data Xm1 .
项17：根据项14或15所述的计算机调色系统或项16所述的预测涂膜的色彩数据的系统，所述系统具备根据已取得的配合组成数据，进行自动调和而实现调色配合的自动调和手段。Item 17: The computer color matching system according to Item 14 or 15 or the system for predicting color data of a coating film according to Item 16, wherein the system has an automatic blending means for automatically blending and achieving color matching based on the obtained matching composition data.
项18：一种应用程序软件，用于对项14～17的任意一项所记载的系统进行控制并使其工作。Item 18: Application software for controlling and operating the system described in any one of Items 14 to 17.
根据本发明，提供一种涂料的制作方法，其用于得到包含光学特性难以预测的光亮性色彩的多种多样色彩的涂色，其不受作业者的熟练度的影响，而是基于通过较少的试制次数就可以结束调色的计算机调色。According to the present invention, a method for preparing a coating is provided, which is used to obtain a variety of colors including bright colors whose optical properties are difficult to predict. The method is not affected by the operator's proficiency and is based on computer color matching that can complete color matching with a small number of trial productions.
根据本发明，提供一种预测涂膜的色彩数据的方法，关于含有光亮性颜料等的多种多样组成的涂料的涂膜，能够高精度地预测其色彩数据。According to the present invention, there is provided a method for predicting color data of a coating film, which can accurately predict the color data of a coating film of a coating material containing a variety of compositions such as a bright pigment.
根据本发明，提供一种计算机调色系统，其用于调制用于取得包含光学特性难以预测的光亮性色彩的多种多样色彩的涂色的涂料，其不受作业者的熟练度的影响，而是通过较少的试制次数就可以结束调色。According to the present invention, a computer color matching system is provided for mixing paint for obtaining a variety of colors including glossy colors whose optical properties are difficult to predict, which is not affected by the operator's proficiency and can complete the color matching with a small number of trial productions.
根据本发明，提供一种预测涂膜的色彩数据的系统，关于含有光亮性颜料等的多种多样组成的涂料的涂膜，能够高精度地预测其色彩数据。According to the present invention, there is provided a system for predicting color data of a coating film, which can predict the color data of a coating film having a variety of compositions including a bright pigment and the like with high accuracy.
由此，通过作业时间的减少或调色次数等的工时削减，能够减轻作业者的负担，通过削减制作的涂料数量，能够实现废弃物的减少及节能化，不受作业者技能的影响而能够实施稳定的调色或色调预测，从而能够得到作业性提高等的效果，工业上极其有用。As a result, the burden on operators can be reduced by reducing operation time or the number of color adjustment times, and waste can be reduced and energy can be saved by reducing the amount of paint produced. Stable color adjustment or hue prediction can be performed without being affected by the operator's skills, thereby achieving effects such as improved workability, which is extremely useful in industry.
附图说明BRIEF DESCRIPTION OF THE DRAWINGS
图1是表示在本发明所涉及的方法中所使用的装置的实施方式的概要化结构图。FIG. 1 is a schematic configuration diagram showing an embodiment of a device used in the method according to the present invention.
图2是表示在本发明所涉及的方法中所使用的装置的其他实施方式的概要化结构图。FIG. 2 is a schematic configuration diagram showing another embodiment of an apparatus used in the method according to the present invention.
图3是在本发明的人工智能模型中的神经网络的概要化结构图。FIG. 3 is a schematic structural diagram of a neural network in the artificial intelligence model of the present invention.
图4是表示本发明的利用多角度分光光度计的变角测色的实施方式的概要化结构图。FIG. 4 is a schematic structural diagram showing an embodiment of the variable angle color measurement using a multi-angle spectrophotometer according to the present invention.
图5是表示本发明的利用多角度分光光度计的变角测色的其他实施方式的概要化结构图。FIG. 5 is a schematic structural diagram showing another embodiment of the variable angle color measurement using a multi-angle spectrophotometer according to the present invention.
图6是表示基于本发明的第1实施形态所涉及的计算机调色的涂料的制作方法的实施方式的流程图。FIG. 6 is a flowchart showing an embodiment of a method for preparing a paint by computer coloring according to the first embodiment of the present invention.
图7表示基于本发明的第2实施形态所涉及的计算机调色的涂料的制作方法的实施方式的流程图。FIG. 7 is a flowchart showing an embodiment of a method for preparing a paint by computer coloring according to a second embodiment of the present invention.
图8是表示本发明的第3实施形态所涉及的预测涂膜的色彩数据的方法的实施方式的流程图。FIG. 8 is a flowchart showing an implementation of a method for predicting color data of a coating film according to a third embodiment of the present invention.
符号说明Explanation of symbols
1-数据库；2、21～24-计算机；31、32-输入装置；41、42-输出装置；51、52-显示装置；61、62-色度计；71、72-摄像仪器；81、82-(自动)调和机；9-神经网络；91-输入层；911-输入节点；92-隐藏层；921-隐藏节点；93-输出层；931-输出节点。1-database; 2, 21-24-computers; 31, 32-input device; 41, 42-output device; 51, 52-display device; 61, 62-colorimeter; 71, 72-camera; 81, 82-(automatic) blender; 9-neural network; 91-input layer; 911-input node; 92-hidden layer; 921-hidden node; 93-output layer; 931-output node.
具体实施方式DETAILED DESCRIPTION
本发明包括：The present invention comprises:
(i)第1实施形态所涉及的涂料的制作方法，其包括生成包含从色彩数据X推断配合组成数据Y的人工智能模型的至少1种的已学习的人工智能模型的工序；(i) A method for producing a coating material according to the first embodiment, comprising the step of generating at least one learned artificial intelligence model including an artificial intelligence model for inferring blending composition data Y from color data X;
(ii)第2实施形态所涉及的涂料的制作方法，其包括生成包含从配合组成数据Y推断色彩数据X的人工智能模型的至少1种的已学习的人工智能模型的工序；(ii) A method for producing a coating material according to the second embodiment, comprising the step of generating at least one learned artificial intelligence model including an artificial intelligence model for inferring color data X from blending composition data Y;
(iii)第3实施形态所涉及的预测涂膜的色彩数据的方法，其包括生成包含从配合组成数据Y推断色彩数据X的人工智能模型的至少1种的已学习的人工智能模型的工序；(iii) A method for predicting color data of a coating film according to a third embodiment, comprising the step of generating at least one learned artificial intelligence model including an artificial intelligence model for inferring color data X from blending composition data Y;
(iv)第4实施形态所涉及的计算机调色系统，其包括生成包含从色彩数据X推断配合组成数据Y的人工智能模型的至少1种的已学习的人工智能模型的手段；(iv) A computer color grading system according to the fourth embodiment, comprising means for generating at least one learned artificial intelligence model including an artificial intelligence model for inferring matching composition data Y from color data X;
(v)第5实施形态所涉及的计算机调色系统，其包括生成包含从配合组成数据Y推断色彩数据X的人工智能模型的至少1种的已学习的人工智能模型的手段；(v) A computer color grading system according to the fifth embodiment, comprising means for generating at least one learned artificial intelligence model including an artificial intelligence model for inferring color data X from combination composition data Y;
(vi)第6实施形态所涉及的预测涂膜的色彩数据的系统，其包括生成包含从配合组成数据Y推断色彩数据X的人工智能模型的至少1种的已学习的人工智能模型的手段；(vi) A system for predicting color data of a coating film according to a sixth embodiment, comprising means for generating at least one learned artificial intelligence model including an artificial intelligence model for inferring color data X from blending composition data Y;
及(vii)应用程序软件，对第4～第6的实施形态所涉及的系统进行控制并使其工作。and (vii) application software for controlling and operating the systems according to the fourth to sixth embodiments.
使用图1～图8，通过以下所例示的实施方式对本发明的方法、系统以及应用程序软件的构成及动作更加详细地进行说明。关于本说明书中的手段及工序，只要能够完成对这些手段及工序进行说明的动作、功能或工序，则并不进行任何限定。另外，只要不脱离本发明的要旨，则本发明并不受以下所例示的实施方式的任何限制。The composition and operation of the method, system and application software of the present invention are described in more detail by using the following illustrative embodiments using Figures 1 to 8. The means and processes in this specification are not limited in any way as long as the actions, functions or processes described by these means and processes can be completed. In addition, the present invention is not limited in any way by the following illustrative embodiments as long as it does not deviate from the gist of the present invention.
装置Device
在基于本发明的计算机调色的涂料的制作方法以及预测涂膜的色彩数据的方法中，使用具备数据库和计算机的装置，在该数据库中至少登记有组合物的色彩数据X及配合组成数据Y，在该计算机中利用登记在该数据库中的数据的配色计算逻辑发挥作用。In the method for preparing a paint using computer color matching and the method for predicting color data of a coating film based on the present invention, a device having a database and a computer is used, in which at least color data X and matching composition data Y of a composition are registered, and in which a color matching calculation logic using the data registered in the database is performed.
在此，所述组合物的色彩数据X包含色彩数据X1～Xn，其分别对应于1种以上的组合物C1～Cn(n为2以上的整数)的各自配合组成数据Y1～Yn。另外，所述组合物的配合组成数据Y包含1种以上的组合物C1～Cn(n为2以上的整数)的各自配合组成数据Y1～Yn。Here, the color data X of the composition includes color data X1 to Xn, which correspond to the respective blending composition data Y1 to Yn of one or more compositions C1 to Cn (n is an integer greater than or equal to 2). In addition, the blending composition data Y of the composition includes the respective blending composition data Y1 to Yn of one or more compositions C1 to Cn (n is an integer greater than or equal to 2).
另外，本发明的计算机调色系统以及预测涂膜的色彩数据的系统具有具备所述数据库和所述使配色计算逻辑发挥作用的计算机的装置。Furthermore, the computer color matching system and the system for predicting color data of a coating film of the present invention include an apparatus including the database and the computer for making the color matching calculation logic function.
图1及图2是表示在基于本发明的计算机调色的涂料的制作方法以及预测涂膜的色彩数据的方法中所使用的装置的实施方式的概要化结构图。该装置还作为本发明的计算机调色系统以及预测涂膜的色彩数据的系统所具备的装置而被使用。Fig. 1 and Fig. 2 are schematic structural diagrams showing an embodiment of a device used in a method for preparing a paint based on computer color matching and a method for predicting color data of a coating film of the present invention. The device is also used as a device possessed by a computer color matching system and a system for predicting color data of a coating film of the present invention.
如图1及图2所示，装置D具备数据库1和计算机2。As shown in FIG. 1 and FIG. 2 , the device D includes a database 1 and a computer 2 .
本发明中，装置D还可以具备2个以上的数据库1，另外，还可以具备2个以上的计算机2。另外，数据库1与计算机2还可以被一体化。In the present invention, the device D may include two or more databases 1, and may include two or more computers 2. In addition, the database 1 and the computer 2 may be integrated.
而且，根据需要，本发明中所使用的装置D还可以具备1个以上的输入装置31(及32)、输出装置41(及42)、显示装置51(及52)、色度计61(及62)、摄像仪器71(及72)、(自动)调和机81(及82)等具有1个以上功能的仪器。另外，还可以1个以上这些仪器与数据库及/或计算机被一体化。Furthermore, if necessary, the device D used in the present invention may also include one or more input devices 31 (and 32), output devices 41 (and 42), display devices 51 (and 52), colorimeters 61 (and 62), camera devices 71 (and 72), (automatic) blenders 81 (and 82) and other devices having one or more functions. In addition, one or more of these devices may be integrated with a database and/or a computer.
在此，数据库1例如可以使用公知的记录装置或服务器等。另外，计算机2可使用市场上销售的个人计算机、便携式终端、智能手机等。Here, for example, a known recording device or server can be used as the database 1. Moreover, as the computer 2, a commercially available personal computer, a portable terminal, a smart phone, etc. can be used.
分别作为输入装置31(及32)而可使用键盘、触摸屏、读取装置等公知的输入装置，作为输出装置41(及42)而可使用印刷装置、数据写入装置等公知的输出装置，作为显示装置51(及52)而可使用显示器等公知的显示装置。As the input device 31 (and 32), a keyboard, a touch screen, a reading device and other well-known input devices can be used; as the output device 41 (and 42), a printing device, a data writing device and other well-known output devices can be used; as the display device 51 (and 52), a display and other well-known display devices can be used.
分别作为色度计61(及62)而可使用(多角度)分光光度计、色彩计、色差计等公知的色度计，作为摄像仪器71(及72)而可使用CCD相机、固态摄像装置、(近)红外线分光摄像装置等公知的摄像装置，作为(自动)调和机81(及82)而可使用包括电子天平装置等的公知的(自动)调和机。As the colorimeter 61 (and 62), a well-known colorimeter such as a (multi-angle) spectrophotometer, a colorimeter, a colorimeter, etc. can be used; as the camera 71 (and 72), a well-known camera device such as a CCD camera, a solid-state camera device, a (near) infrared spectroscopic camera device can be used; as the (automatic) blender 81 (and 82), a well-known (automatic) blender including an electronic balance device can be used.
在本发明中所使用的装置中，数据库、计算机及所述仪器通过有线、无线或组合这些的通信手段或借由记录介质的手段，被连接成相互可发收数据。作为通信手段例如可例举LAN(局域网)、WAN(广域网络)、互联网、电话网等各种通信网络的1个以上的组合。In the device used in the present invention, the database, computer and the instrument are connected to each other by wired, wireless or a combination of these communication means or by means of a recording medium so that data can be sent and received. Examples of the communication means include a combination of one or more of various communication networks such as LAN (local area network), WAN (wide area network), the Internet, and a telephone network.
图1是1个数据库1与1个计算机2以相互可发收数据的方式被连接的装置的例子。数据库1上还可以连接有输入装置31、输出装置41、显示装置51、色度计61、摄像仪器71及(自动)调和机81的任意1个以上的仪器，而且计算机2上也还可以连接有输入装置32、输出装置42、显示装置52、色度计62、摄像仪器72及(自动)调和机82的任意1个以上的仪器。连接于数据库1或计算机2的色度计61、62及摄像仪器71、72及(自动)调和机81、82的1个以上根据来自数据库1或计算机2的指令进行测定或调和等。另外，将测定出的数据等发送至数据库1或计算机2，最终能够将数据登记在数据库1中。FIG. 1 is an example of a device in which a database 1 and a computer 2 are connected in a manner that data can be sent and received to each other. The database 1 may be connected to any one or more of an input device 31, an output device 41, a display device 51, a colorimeter 61, a camera 71, and an (automatic) blender 81, and the computer 2 may be connected to any one or more of an input device 32, an output device 42, a display device 52, a colorimeter 62, a camera 72, and an (automatic) blender 82. One or more of the colorimeters 61, 62, the cameras 71, 72, and the (automatic) blenders 81, 82 connected to the database 1 or the computer 2 perform measurement or blending according to instructions from the database 1 or the computer 2. In addition, the measured data is sent to the database 1 or the computer 2, and the data can be finally registered in the database 1.
在图1的装置中，数据库1还可以形成于计算机2内的记录装置，此时，不需要进行与数据库1的通信，而是该装置能够单独独立进行计算机调色及涂膜的色彩数据的预测。另外，根据需要，能够在适当的时刻对登记在数据库中的数据进行更新(update)来维护数据库，由此作业者能够根据最新的数据进行作业。In the device of FIG. 1 , the database 1 can also be formed as a recording device in the computer 2. In this case, there is no need to communicate with the database 1, but the device can independently perform computer color matching and prediction of color data of the coating film. In addition, as needed, the data registered in the database can be updated at an appropriate time to maintain the database, so that the operator can work according to the latest data.
图2是对1个数据库1连接有2个以上的计算机21～2X的装置的例子。图2中，示出了连接有4个计算机21～24的例子。数据库1还可以构成为2个以上的数据库1可通信地被连接。通过增加数据库1的数量，能够增加可连接成可通信的计算机的数量。FIG. 2 shows an example of a device in which two or more computers 21 to 2X are connected to one database 1. FIG. 2 shows an example in which four computers 21 to 24 are connected. The database 1 may also be configured such that two or more databases 1 are connected so as to be communicable. By increasing the number of databases 1, the number of computers that can be connected so as to be communicable can be increased.
图2中，数据库1上既可以连接有输入装置31、输出装置41及显示装置51，还可以连接有色度计、摄像仪器及(自动)调和机的任意一个以上的仪器。In FIG. 2 , the database 1 may be connected to an input device 31 , an output device 41 , and a display device 51 , and may also be connected to any one or more of a colorimeter, a camera, and an (automatic) blender.
图2中，计算机21～24上还可以连接有输入装置32、输出装置42、显示装置52、色度计62、摄像仪器72及(自动)调和机82的任意一个以上的仪器。In FIG. 2 , the computers 21 to 24 may be connected to any one or more of an input device 32 , an output device 42 , a display device 52 , a colorimeter 62 , a camera 72 and an (automatic) blender 82 .
图2的装置D相当于将数据库1作为服务器而连接有多个计算机2。例如，还可以构成为，在由涂料公司等进行管理的数据库1上借由通信线路(例如，互联网线路、电话线路等)连接有作业者(用户)的计算机2，由此可进行数据通信。The device D in Fig. 2 is equivalent to a database 1 as a server connected to a plurality of computers 2. For example, it is also possible to connect a computer 2 of an operator (user) to a database 1 managed by a paint company or the like via a communication line (e.g., an Internet line, a telephone line, etc.), thereby enabling data communication.
图3是表示从色彩数据X推断配合组成数据Y的人工智能模型或从配合组成数据Y推断色彩数据X的人工智能模型中的神经网络9(用程序再现大脑的神经细胞的活动)的概要化结构图。3 is a schematic structural diagram showing a neural network 9 (a program that reproduces the activity of nerve cells in the brain) in an artificial intelligence model that infers matching composition data Y from color data X or infers color data X from matching composition data Y.
使用登记在数据库中的数据，在向计算机输入学习用数据的同时使学习用数据进行机器学习，由此生成人工智能模型。如图3所示，神经网络9构成为包含输入层91、隐藏层92及输出层93这3个处理层(3个神经元层)。Using the data registered in the database, the learning data is input into the computer and the learning data is subjected to machine learning, thereby generating an artificial intelligence model. As shown in FIG3 , the neural network 9 is composed of three processing layers (three neuron layers) including an input layer 91 , a hidden layer 92 and an output layer 93 .
输入层91包含称为输入节点911～91i的至少1～i个处理要素，接合于网络的隐藏层92的隐藏节点921～92j。The input layer 91 includes at least 1 to i processing elements, referred to as input nodes 911 to 91i, which are connected to hidden nodes 921 to 92j of the hidden layer 92 of the network.
在本发明的第1实施形态所涉及的涂料的制作方法以及本发明的第4实施形态所涉及的计算机调色系统中，所使用的在从色彩数据X推断配合组成数据Y的人工智能模型中的神经网络9的输入层91的各单元，对应于色彩数据X所涉及的1种以上的各特征量。In the paint production method involved in the first embodiment of the present invention and the computer color matching system involved in the fourth embodiment of the present invention, each unit of the input layer 91 of the neural network 9 used in the artificial intelligence model for inferring the combination composition data Y from the color data X corresponds to one or more feature quantities involved in the color data X.
在本发明的第2实施形态所涉及的涂料的制作方法、本发明的第3实施形态所涉及的预测涂膜的色彩数据的方法、本发明的第4实施形态所涉及的计算机调色系统以及本发明的第6实施形态所涉及的预测涂膜的色彩数据的系统中，所使用的在从配合组成数据Y推断色彩数据X的人工智能模型中的神经网络9的输入层91的各单元，对应于配合组成数据Y所涉及的1种以上的各特征量。In the method for preparing paint involved in the second embodiment of the present invention, the method for predicting color data of a coating film involved in the third embodiment of the present invention, the computer color matching system involved in the fourth embodiment of the present invention, and the system for predicting color data of a coating film involved in the sixth embodiment of the present invention, each unit of the input layer 91 of the neural network 9 in the artificial intelligence model for inferring color data X from the combination composition data Y corresponds to one or more feature quantities involved in the combination composition data Y.
隐藏层92具有称为隐藏节点921～92j的至少1～j个处理要素，接合于网络的输出层93的输出节点931～93k。隐藏层92(隐藏节点921～92j)存在于输入层91(输入节点911～91i)与输出层93(输出节点931～93k)之间。为了对输入输出关系的复杂程度进行模型化，能够通过对追加于网络功能的隐藏节点的数量进行增减来改变隐藏节点921～92j的数量。The hidden layer 92 has at least 1 to j processing elements called hidden nodes 921 to 92j, and is connected to the output nodes 931 to 93k of the output layer 93 of the network. The hidden layer 92 (hidden nodes 921 to 92j) exists between the input layer 91 (input nodes 911 to 91i) and the output layer 93 (output nodes 931 to 93k). In order to model the complexity of the input-output relationship, the number of hidden nodes 921 to 92j can be changed by increasing or decreasing the number of hidden nodes added to the network function.
输出层93被组织成具有称为输出节点931～93k的至少1～k个处理要素。处理要素或节点被相互接合，以便在网络运行时能够计算出配合组成数据与色彩数据之间的关系。The output layer 93 is organized into at least 1 to k processing elements, referred to as output nodes 931 to 93k. The processing elements or nodes are interconnected so that the relationship between the coordinated component data and the color data can be calculated when the network is running.
本发明的第1实施形态所涉及的涂料的制作方法以及本发明的第4实施形态所涉及的计算机调色系统中，所使用的在从色彩数据X推断配合组成数据Y的人工智能模型中的神经网络9的输出层93的各单元，对应于配合组成数据Y所涉及的1种以上的各特征量。In the paint production method involved in the first embodiment of the present invention and the computer color matching system involved in the fourth embodiment of the present invention, each unit of the output layer 93 of the neural network 9 in the artificial intelligence model for inferring the combination composition data Y from the color data X corresponds to one or more feature quantities involved in the combination composition data Y.
在本发明的第2实施形态所涉及的涂料的制作方法、本发明的第3实施形态所涉及的预测涂膜的色彩数据的方法、本发明的第4实施形态所涉及的计算机调色系统以及本发明的第6实施形态所涉及的预测涂膜的色彩数据的系统中，所使用的在从配合组成数据Y推断色彩数据X的人工智能模型中的神经网络9的输出层93的各单元，对应于色彩数据X所涉及的1种以上的各特征量。In the method for preparing a paint involved in the second embodiment of the present invention, the method for predicting color data of a coating involved in the third embodiment of the present invention, the computer color matching system involved in the fourth embodiment of the present invention, and the system for predicting color data of a coating involved in the sixth embodiment of the present invention, each unit of the output layer 93 of the neural network 9 in the artificial intelligence model for inferring color data X from the combination composition data Y corresponds to one or more feature quantities involved in the color data X.
并且，神经网络9内数据只在1个方向上流动，各节点不会将信号只向1个以上的节点发送就接收反馈。Furthermore, data in the neural network 9 flows in only one direction, and each node will not send a signal to only one or more nodes and then receive feedback.
在本发明的第1实施形态所涉及的涂料的制作方法以及本发明的第4实施形态所涉及的计算机调色系统中，所使用的在从色彩数据X推断配合组成数据Y的人工智能模型中，输入层91中的输入节点911～91i，对于各色彩数据中的1的输入变量(输入要素；参数)，对应于1个输入节点。另外，输出层93中的输出节点931～93k，对于各配合组成数据中的1个输出变量(输出要素；参数)，对应于1个输出节点。In the artificial intelligence model for inferring the combination composition data Y from the color data X used in the paint production method according to the first embodiment of the present invention and the computer color matching system according to the fourth embodiment of the present invention, the input nodes 911 to 91i in the input layer 91 correspond to one input node for one input variable (input element; parameter) in each color data. In addition, the output nodes 931 to 93k in the output layer 93 correspond to one output node for one output variable (output element; parameter) in each combination composition data.
在本发明的第2实施形态所涉及的涂料的制作方法、本发明的第3实施形态所涉及的预测涂膜的色彩数据的方法、本发明的第4实施形态所涉及的计算机调色系统以及本发明的第6实施形态所涉及的预测涂膜的色彩数据的系统中，所使用的在从配合组成数据Y推断色彩数据X的人工智能模型中，输入层91中的输入节点911～91i，对于各配合组成数据中的1个输入变量(输入要素；参数)，对应于1个输入节点。另外，输出层93中的输出节点931～93k，对于各色彩数据中的1个输出变量(输出要素；参数)，对应于1个输出节点。In the artificial intelligence model for inferring color data X from combination composition data Y used in the coating preparation method according to the second embodiment of the present invention, the method for predicting color data of a coating film according to the third embodiment of the present invention, the computer color matching system according to the fourth embodiment of the present invention, and the system for predicting color data of a coating film according to the sixth embodiment of the present invention, the input nodes 911 to 91i in the input layer 91 correspond to one input node for one input variable (input element; parameter) in each combination composition data. In addition, the output nodes 931 to 93k in the output layer 93 correspond to one output node for one output variable (output element; parameter) in each color data.
人工智能模型中，隐藏层92中的隐藏节点921～92j的数量，对应于输入输出关系的复杂程度而能够发生增减。输入层91的各输入节点间、输入节点与隐藏节点间、隐藏节点与输出节点间、输出层93的各输出节点间的各自接合具有与此有关的接合加权，而且分别对隐藏节点921～92j及输出节点931～93k，还可以具有一个以上的追加阈值加权。In the artificial intelligence model, the number of hidden nodes 921-92j in the hidden layer 92 can be increased or decreased according to the complexity of the input-output relationship. Each connection between the input nodes of the input layer 91, between the input nodes and the hidden nodes, between the hidden nodes and the output nodes, and between the output nodes of the output layer 93 has a connection weight associated therewith, and each hidden node 921-92j and output node 931-93k can also have one or more additional threshold weights.
在分析复杂且因用于计算机专门系统中而从人的知识中导出模型比较麻烦的复杂的系统或现象中，使用神经网络9的人工智能尤其有利。Artificial intelligence using neural networks 9 is particularly advantageous in analyzing complex systems or phenomena that are difficult to analyze and for which it is troublesome to derive models from human knowledge for use in computer-specific systems.
并且，例如在图2所示的装置中，神经网络9可生成在构成数据库的服务器计算机侧。由此，已连接的作业者(用户)不被地理性重要条件等所左右，而是随时可接收高品质的数据的提供。另外，通过提高服务器计算机的安全，能够防止缘于不正确连接的数据改变及神经网络的破坏等。Furthermore, for example, in the device shown in FIG2 , the neural network 9 can be generated on the server computer side constituting the database. Thus, the connected operator (user) can receive high-quality data at any time without being affected by geographically important conditions, etc. In addition, by improving the security of the server computer, data changes due to improper connection and destruction of the neural network can be prevented.
本发明中，在从色彩数据X推断配合组成数据Y的人工智能模型的至少1种的基础上，还可以具备从配合组成数据Y推断色彩数据X的人工智能模型的至少1种。在该人工智能模型中的神经网络构成为与图3所涉及的神经网络相同，分别输入层91的各单元对应于配合组成数据Y所涉及的1种以上的各特征量，输出层93的各单元对应于色彩数据X所涉及的1种以上的各特征量。In the present invention, in addition to at least one artificial intelligence model for inferring the matching composition data Y from the color data X, at least one artificial intelligence model for inferring the color data X from the matching composition data Y may be provided. The neural network in the artificial intelligence model is configured similarly to the neural network involved in FIG. 3 , and each unit of the input layer 91 corresponds to each of more than one feature quantity involved in the matching composition data Y, and each unit of the output layer 93 corresponds to each of more than one feature quantity involved in the color data X.
数据库database
本发明中将色彩数据与配合组成数据对应起来进行登记(记录)，由此构成数据库。在本发明的数据库中，在色彩数据及配合组成数据的基础上，还可以对应起来登记与色彩或组成有关的各种数据。本发明中，登记在数据库中的数据优选非常多的数据(所谓大数据)。具体而言是5千组以上，优选1万组以上，更优选2万组以上。可任意地追加、变更、消除这些数据。In the present invention, color data and matching composition data are registered (recorded) in correspondence, thereby forming a database. In the database of the present invention, various data related to color or composition can also be registered in correspondence on the basis of color data and matching composition data. In the present invention, the data registered in the database is preferably a very large amount of data (so-called big data). Specifically, it is more than 5,000 groups, preferably more than 10,000 groups, and more preferably more than 20,000 groups. These data can be added, changed, and deleted at will.
登记在数据库中的1种以上的组合物的配合组成数据Y及对应的色彩数据X包含实测数据或者包含实测数据和根据实测数据算出的数据。作为根据实测数据算出的数据，可例举使用实测数据通过规定的公式等算出的各种参数值及根据实测数据可算出的预测值等。The compounding composition data Y of one or more compositions registered in the database and the corresponding color data X include measured data or measured data and data calculated from the measured data. Examples of the data calculated from the measured data include various parameter values calculated by a predetermined formula using the measured data and predicted values that can be calculated from the measured data.
数据库既可以是单一的数据库，另外还可以是至少通过一个共通信息要素相互关联起来的多个数据库或者并未关联起来的多个数据库。The database may be a single database, or may be a plurality of databases that are linked to each other via at least one common information element, or a plurality of databases that are not linked to each other.
数据库可设置在可进行与计算机的通信且可遥控操作的服务器中。另外，数据库的至少一部分可设置在计算机、色度计、微观光亮感测定装置、自动调和机以及取得或使用登记在其他数据库中的数据的装置中的记录部(存储器等)。The database may be provided in a server that can communicate with a computer and can be remotely controlled. In addition, at least a part of the database may be provided in a recording unit (memory, etc.) in a computer, a colorimeter, a microscopic light perception measuring device, an automatic blender, or a device that obtains or uses data registered in another database.
当数据库为多个时，各数据库还可以通过有线或无线连接。另外，数据库还可以通过有线或无线连接于计算机、色度计、微观光亮感测定装置、自动调和机以及取得或使用登记在其他数据库中的数据的装置的一个以上。When there are multiple databases, each database can also be connected by wire or wireless. In addition, the database can also be connected by wire or wireless to one or more of a computer, a colorimeter, a microscopic light perception measuring device, an automatic blender, and a device for obtaining or using data registered in other databases.
组合物Composition
本发明中的组合物还可以以任意量比含有1种以上的色料。包含于组合物中的色料例如为着色颜料、染料、光亮性颜料(光反射性颜料、光干涉性颜料等)等具有使组合物发色的功能的材料。The composition of the present invention may also contain one or more colorants in any amount ratio. The colorants contained in the composition are, for example, coloring pigments, dyes, bright pigments (light reflecting pigments, light interfering pigments, etc.), and other materials that have the function of making the composition develop color.
本发明的组合物例如还可以使用修补时被使用的“原色涂料”。另外，例如还可以是对含有1个以上的原色涂料的原材料进行混合而配色成所希望的颜色的涂料。The composition of the present invention may be a "primary color paint" used for repairing, for example. In addition, for example, raw materials containing one or more primary color paints may be mixed to match the color to a desired color.
本发明的组合物还可以含有着色颜料浆料、光亮性颜料浆料、定向控制剂、光泽控制剂以及其他涂料领域等中使用的各种添加剂等。The composition of the present invention may further contain a coloring pigment slurry, a bright pigment slurry, an orientation control agent, a gloss control agent, and various additives used in other coating fields.
色彩数据Color Data
本发明中，登记在数据库中的色彩数据包括颜色所涉及的数据以及质感、光亮感、光泽等外观特性所涉及的数据。In the present invention, the color data registered in the database includes data related to color and data related to appearance characteristics such as texture, gloss, and luster.
可使用色度计或摄像仪器等仪器对从所述组合物得到的涂膜进行测定而取得这些数据。另外，还可以通过使用仪器取得的涂膜的1个以上的图像数据或根据需要对该图像数据进行解析、变换、修正等而取得。而且，将通过对图像数据进行处理而得到的通过测定得到的各种色彩数据的至少一部分，还可以进行运算处理而算出。并且，使用仪器进行测定而得到的数据还可以是，根据需要对起因于测定仪器间或测定变动等的误差等进行修正的数据。These data can be obtained by measuring the coating obtained from the composition using instruments such as a colorimeter or a camera. In addition, the coating obtained using an instrument can be obtained by analyzing, converting, correcting, etc., or by analyzing, converting, or correcting the image data as needed. Moreover, at least a portion of the various color data obtained by measuring and processing the image data can also be calculated by arithmetic processing. Furthermore, the data obtained by measuring using an instrument can also be data corrected for errors caused by measuring instruments or measuring changes, etc. as needed.
另外，还可以将组合物自身的K值(光吸收系数)及S值(光散射系数)作为色彩数据而加以使用。K值及S值例如能够对组合物及减淡组合物的颜色的测色数据进行数值处理而得到。In addition, the K value (light absorption coefficient) and S value (light scattering coefficient) of the composition itself can also be used as color data. The K value and S value can be obtained by numerically processing the color measurement data of the composition and the lightening composition, for example.
颜色所涉及的数据及/或外观特性所涉及的数据例如为使用色彩计、多角度分光光度计、激光式金属感测定仪器、变角分光光度计、光泽计、摄像仪器、微观光亮感测定仪等测定仪器的测定直接取得的数据，或者从通过测定取得的数据算出的数据。Data related to color and/or data related to appearance characteristics are, for example, data directly obtained by measurement using measuring instruments such as a colorimeter, a multi-angle spectrophotometer, a laser metallicity measuring instrument, a variable angle spectrophotometer, a glossmeter, a camera, a microscopic glossiness measuring instrument, or data calculated from data obtained through measurement.
另外，颜色所涉及的数据及/或外观特性所涉及的数据可包含1个或多个照明角度、1个或多个观察角度或与这些组合有关联的图像等数据。In addition, the data related to color and/or the data related to appearance characteristics may include data such as one or more lighting angles, one or more observation angles, or images related to these combinations.
作为用于取得色彩数据的计量仪器，只要是可测定光亮涂膜(金属涂膜、珍珠色涂膜等)、纯色涂膜等的色彩而取得色彩数据的计量仪器，则并不特意限制采用测定原理、测定值的色彩数据的算出方法等，而是可使用现有公知的计量仪器。例如，可使用具备照射被测色表面的光源的单角度分光光度计、多角度分光光度计、色彩计、色差计、变角分光光度计等色度计以及摄像装置、微观光亮感测定仪等仪器以及色样本卡等计量仪器的1个以上。另外，可任意使用对从这些计量仪器得到的各种色彩数据进行处理的数据处理装置。As a measuring instrument for obtaining color data, as long as it can measure the color of a bright coating (metallic coating, pearl coating, etc.), a solid color coating, etc. and obtain color data, there is no particular limitation on the measurement principle adopted, the method of calculating the color data of the measured value, etc., and any existing known measuring instrument can be used. For example, a colorimeter such as a single-angle spectrophotometer, a multi-angle spectrophotometer, a colorimeter, a colorimeter, a variable-angle spectrophotometer, etc., which has a light source for irradiating the color surface to be measured, and an imaging device, a microscopic brightness measurement instrument, and a color sample card can be used. In addition, a data processing device that processes various color data obtained from these measuring instruments can be used arbitrarily.
登记在数据库中的色彩数据中，作为颜色所涉及的数据，可例举表示亮度、彩度、色相的数据或通过计算可确定颜色的数据。例如，可以是基于XYZ表色系(X、Y、Z值)、RGB表色系、L＊a＊b＊表色系(L＊、a＊、b＊值)、HunterLab表色系(L、a、b值)、由CIE(1994)规定的L＊C＊h表色系(L＊、C＊、h值)、孟塞尔表色系(H、V、C值)等表色系的数据。本发明中，登记在数据库中的色彩数据中的颜色所涉及的数据，可以是基于这些中的任意1个以上的表色系的数据。优选为在包括修补涂装领域的各种领域中广泛使用的基于L＊a＊b＊表色系或L＊C＊h表色系的数据。In the color data registered in the database, as data related to the color, data representing brightness, chroma, hue or data that can determine the color by calculation can be cited. For example, it can be data based on color systems such as the XYZ color system (X, Y, Z values), the RGB color system, the L*a*b* color system (L*, a*, b* values), the HunterLab color system (L, a, b values), the L*C*h color system (L*, C*, h values) specified by CIE (1994), and the Munsell color system (H, V, C values). In the present invention, the data related to the color in the color data registered in the database can be data based on any one or more of these color systems. Preferably, it is data based on the L*a*b* color system or the L*C*h color system that is widely used in various fields including the field of repair painting.
登记在数据库中的色彩数据中，作为外观特性所涉及的数据，例如可例举当观察含有光亮性颜料的涂膜等被测色表面时感觉到的质感的宏观观察下可察觉的质感即宏观光亮感以及微观观察下可察觉的质感即微观光亮感、纵深感觉(深度感)、鲜明度等。Among the color data registered in the database, data related to appearance characteristics include, for example, the texture felt when observing the surface of the measured color such as a coating containing a bright pigment, the texture perceptible under macroscopic observation, i.e., macroscopic brightness, and the texture perceptible under microscopic observation, i.e., microscopic brightness, depth perception, vividness, etc.
宏观光亮感Macro light sensation
作为宏观光亮感，可例举向被测色表面照射均匀的光，对反射的光以每个角度受光而测定颜色来得到的多角度分光反射率。另外，当远距离观察被测色表面时，可例举表示因照明与观察角度的增减而颜色(亮度、彩度、色相)发生变化的触发现象的FF值(触发值)以及表示在射入光的相反侧出现的与正反射光之间的开度角为10度至25度之间的强光侧的目视亮度的IV值(intensity value值)以及表示强光侧的正面的亮度的SV值(scattervalue值)、cFF值、金属感指数、深度感指数、鲜明度、表示光泽的光泽值等。As a macroscopic glossiness, the multi-angle spectral reflectance obtained by irradiating the measured color surface with uniform light and measuring the color of the reflected light at each angle can be cited. In addition, when the measured color surface is observed from a distance, the FF value (trigger value) indicating the triggering phenomenon of the color (brightness, chroma, hue) changing due to the increase or decrease of the illumination and observation angle, the IV value (intensity value) indicating the visual brightness of the strong light side with an opening angle between 10 and 25 degrees between the regular reflected light and the opposite side of the incident light, the SV value (scatter value) indicating the brightness of the front side of the strong light side, the cFF value, the metallic index, the depth index, the sharpness, the gloss value indicating the gloss, etc. can be cited.
宏观光亮感例如通过多角度分光光度计、激光式金属感测定仪器、变角分光光度计、光泽计等可直接取得，另外还可以从这些算出。例如，作为多角度分光光度计，可使用BYK-Mac i(商品名、BYK公司制)、MA-68II(商品名、X-Rite公司制)等。通过使用激光式金属感测定装置ALCOPE(注册商标)LMR-200(商品名、关西涂料公司制)，还可以求出FF值、IV值及SV值。The macroscopic glossiness can be directly obtained, for example, by a multi-angle spectrophotometer, a laser metal-sensitivity measuring instrument, a variable angle spectrophotometer, a gloss meter, etc., and can also be calculated from these. For example, BYK-Mac i (trade name, manufactured by BYK), MA-68II (trade name, manufactured by X-Rite) and the like can be used as a multi-angle spectrophotometer. By using a laser metal-sensitivity measuring device ALCOPE (registered trademark) LMR-200 (trade name, manufactured by Kansai Paint Co., Ltd.), FF value, IV value and SV value can also be obtained.
多角度分光反射率是用可多角度测色的分光光度计进行测定的分光反射率且用R(x、λ)表示。在此，R是分光反射率(Reflectance)，用通过附属于测定仪的校正板进行校正的分光反射率％表示。X是受光角度，用与正反射光之间的偏角表示。λ是波长，以10nm间隔(波长数31个)测定了可见光范围400～700nm。作为入射角度，是作为通常标准的-45度。The multi-angle spectral reflectance is the spectral reflectance measured by a spectrophotometer capable of multi-angle color measurement and is represented by R(x, λ). Here, R is the spectral reflectance (Reflectance), which is represented by the spectral reflectance % corrected by the correction plate attached to the measuring instrument. X is the angle of light reception, which is represented by the deflection angle between the regular reflected light. λ is the wavelength, and the visible light range of 400 to 700nm is measured at 10nm intervals (31 wavelengths). As the angle of incidence, -45 degrees is the usual standard.
本发明中，在将入射角度做成45度时，受光角度x是从强光(high-light)(25度、15度、-15度)、正面(face)(45度)到遮光面(shade)(75度、110度)的任意1个角度以上，优选3个角度以上。如图4所示，优选受光角度为-15度、15度、25度、45度、75度及110度的6个角度。另外，如图5所示，受光角度还可以为15度、25度、45度、75度及110度的5个角度。In the present invention, when the incident angle is made into 45 degrees, the light receiving angle x is any one angle or more from high-light (25 degrees, 15 degrees, -15 degrees), front (face) (45 degrees) to shade (75 degrees, 110 degrees), preferably three angles or more. As shown in Figure 4, the light receiving angle is preferably 6 angles of -15 degrees, 15 degrees, 25 degrees, 45 degrees, 75 degrees and 110 degrees. In addition, as shown in Figure 5, the light receiving angle can also be 5 angles of 15 degrees, 25 degrees, 45 degrees, 75 degrees and 110 degrees.
FF值(触发值)是表示基于观察角度(受光角)的L值(亮度)的变化程度的值。触发则对应于强光方向(靠近光的正反射方向的方向)与遮光面方向(远离光的正反射方向的方向)上的亮度差。FF值越大，则基于观察角度(受光角)的L值(亮度)的变化越大，表示触发性出色。The FF value (trigger value) is a value that indicates the degree of change in the L value (brightness) based on the observation angle (light receiving angle). The trigger corresponds to the brightness difference between the strong light direction (the direction close to the regular reflection direction of the light) and the light shielding surface direction (the direction away from the regular reflection direction of the light). The larger the FF value, the greater the change in the L value (brightness) based on the observation angle (light receiving angle), indicating excellent triggering.
在被测色表面上，使用多角度分光色度计从45度的角度照射光，测定受光角15度及受光角110度的L值(亮度)，通过下式可求出FF值。On the color surface to be measured, use a multi-angle spectrophotometer to irradiate light from an angle of 45 degrees, measure the L value (brightness) at an angle of 15 degrees and an angle of 110 degrees, and calculate the FF value using the following formula.
FF值＝受光角15度的L值/受光角110度的L值FF value = L value at an acceptance angle of 15 degrees / L value at an acceptance angle of 110 degrees
IV值是从偏角为15度的方向上测定出的分光反射率求出的XYZ表色系上的Y值。The IV value is the Y value on the XYZ colorimetric system obtained from the spectral reflectance measured in a direction with an off-angle of 15 degrees.
SV值是从偏角为45度的方向上测定出的分光反射率求出的XYZ表色系上的Y值。The SV value is the Y value on the XYZ colorimetric system obtained from the spectral reflectance measured in a direction with an off-angle of 45 degrees.
将从偏角为15度的方向上测定出的分光反射率求出的XYZ表色系上的Y值作为Y15，将从偏角为45度的方向上测定出的分光反射率求出的XYZ表色系上的Y值作为Y45，通过下式可求出FF值。The FF value can be calculated by the following formula: Y15 is the Y value on the XYZ colorimetric system obtained from the spectral reflectance measured in the direction with an offset angle of 15 degrees, and Y45 is the Y value on the XYZ colorimetric system obtained from the spectral reflectance measured in the direction with an offset angle of 45 degrees.
FF值＝2×(Y15-Y45)/(Y15+Y45)FF value = 2 × (Y15-Y45) / (Y15+Y45)
将从偏角为15度的方向上测定出的分光反射率求出的L＊c＊h表色系上的c＊值作为c*15，将从偏角为45度的方向上测定出的分光反射率求出的L＊c＊h表色系上的c＊作为c*45，通过下式可求出cFF值。The cFF value can be calculated by the following formula: the c* value on the L*c*h colorimetric system obtained from the spectral reflectance measured in the direction with a deviation angle of 15 degrees is taken as c*15, and the c* value on the L*c*h colorimetric system obtained from the spectral reflectance measured in the direction with a deviation angle of 45 degrees is taken as c*45.
cFF值＝2×(c*15-c*45)/(c*15+c*45)cFF value = 2×(c*15-c*45)/(c*15+c*45)
将从所述偏角为15度的方向上测定出的分光反射率求出的XYZ表色系上的Y值作为Y15，将所述FF值作为FF，通过下式可求出金属感指数。The metallic index can be calculated by the following formula by taking the Y value on the XYZ colorimetric system obtained from the spectral reflectance measured in the direction with the off-angle of 15 degrees as Y15 and the FF value as FF.
金属感指数＝Y15×FF2Metallic index = Y15 × FF2
深度感指数表示由光亮性颜料赋予的纵深感觉，将从特性角度下的分光反射率求出的L＊c＊h表色系上的L＊值及c＊值分别作为L*R及c*R，使用这些可用下式求出。The depth index represents the depth sensation imparted by the bright pigment. The L* value and c* value on the L*c*h colorimetric system obtained from the spectral reflectance at a characteristic angle are respectively referred to as L*R and c*R. The depth index can be obtained by the following formula.
深度感指数＝c*R/L*RDepth Perception Index = c*R/L*R
例如，使用多角度分光色度计从45度的角度照射光，测定受光角75度的c＊值即c*75及受光角75度的L＊值即L*75，可使用通过下式求出的遮光面的深度感。For example, by irradiating light from an angle of 45 degrees using a multi-angle spectrophotometer and measuring the c* value c*75 at a light receiving angle of 75 degrees and the L* value L*75 at a light receiving angle of 75 degrees, the depth of the light shielding surface can be obtained using the following formula.
遮光面的深度感＝c*75/L*75Depth of shading surface = c*75/L*75
使用所述L*R及c*R，通过下式可求出鲜明度。Using the above L*R and c*R, the clarity can be calculated by the following formula.
鲜明度＝sqrt[(L*R)2+(c*R)2]Sharpness = sqrt[(L*R)2 +(c*R)2 ]
微观光亮感Microscopic light
微观光亮感是，当近距离观察被测色表面时，因被测色表面中的光亮性颜料而呈现的如闪亮感或粒子感这样的关于二维性的亮度不均匀的感觉。Microscopic glossiness is a feeling of uneven brightness in two dimensions, such as sparkle or particles, caused by the glossy pigments on the surface of the color being measured when the surface is observed at a close distance.
可使用微观光亮感测定仪取得微观光亮感。作为微观光亮感测定仪，例如可例举具备向光亮涂膜面照射光的光照射装置及以照射光并入射入的角度对被光照射的涂膜面进行拍摄而形成图像的CCD相机及连接于该CCD相机的解析该图像的图像解析装置的微观光亮感测定仪等。The microscopic light perception can be obtained by using a microscopic light perception meter. Examples of the microscopic light perception meter include a light irradiation device that irradiates light to a glossy coating surface, a CCD camera that photographs the irradiated coating surface at an angle of incidence of the irradiated light to form an image, and an image analysis device that analyzes the image and is connected to the CCD camera.
当使用微观光亮感测定仪测定被测色表面的微观光亮感时，首先向被测色表面照射模拟(人工)太阳光。相对于被测色表面的铅直线，向被测色表面的光照射角度通常为5～60度，优选10～20度的范围内比较适合，尤其适合相对于铅直线呈15度左右。另外，虽然并不特意限定光的照射区域的形状，但是通常为圆形，虽然被测色表面上的照射面积通常适合被测色表面的1～10，000mm2的范围内，但是并不限制该范围。照射光的照度通常优选100～2，000勒克斯(lux)的范围内。When using a microscopic brightness meter to measure the microscopic brightness of the color surface being measured, first irradiate the color surface being measured with simulated (artificial) sunlight. The light irradiation angle to the color surface being measured is usually 5 to 60 degrees relative to the vertical line of the color surface being measured, preferably 10 to 20 degrees, and especially about 15 degrees relative to the vertical line. In addition, although the shape of the light irradiation area is not specifically limited, it is usually circular. Although the irradiation area on the color surface being measured is usually suitable for the range of 1 to 10,000mm2 of the color surface being measured, this range is not limited. The illumination of the irradiated light is usually preferably in the range of 100 to 2,000 lux.
向被测色表面照射光，在基于此的反射光中以正反射光并不射入的角度用CCD(Charge Coupled Device)相机拍摄被光照射的被测色表面。虽然该拍摄角度为正反射光并不射入的角度即可，但是相对于被测色表面尤其适合铅直方向。另外，优选CCD相机的拍摄方向与正反射光的角度处于10～60度的范围内。只要在被光照射的被测色表面上的CCD相机的测定范围为均匀地被光照射的范围，则并不特意进行限定，但是通常包括照射部分的中央部，测定面积比较适合1～10，000mm2，优选10～600mm2の范围内比较适合。The color surface to be measured is irradiated with light, and the color surface to be measured that is irradiated with light is photographed with a CCD (Charge Coupled Device) camera at an angle at which regular reflected light does not enter in the reflected light based on the light. Although the shooting angle may be an angle at which regular reflected light does not enter, it is particularly suitable for the vertical direction relative to the color surface to be measured. In addition, it is preferred that the shooting direction of the CCD camera and the angle of regular reflected light are within the range of 10 to 60 degrees. As long as the measurement range of the CCD camera on the color surface to be measured that is irradiated with light is a range that is uniformly irradiated with light, it is not particularly limited, but it usually includes the central part of the irradiated part, and the measurement area is more suitable for 1 to 10,000mm2 , preferably 10 to 600mm2 .
用上述CCD相机拍摄的图像是2维图像，被分割成多个(通常10，000～1，000，000个)区域(像点、像素)，测定各个区域中的亮度。本发明中，“亮度”是表示“用CCD相机拍摄得到的2维图像的每一个区域的浓淡值的数码等级，对应于被摄体的亮度的数码量”。数字等级表示从8二进制位分解度的CCD相机输出的每一个区域的亮度，用0～255的值表示该数码等级。The image captured by the above-mentioned CCD camera is a two-dimensional image, which is divided into a plurality of (usually 10,000 to 1,000,000) regions (pixels), and the brightness in each region is measured. In the present invention, "brightness" means "the digital level of the shading value of each region of the two-dimensional image captured by the CCD camera, corresponding to the digital amount of the brightness of the subject". The digital level represents the brightness of each region output from the CCD camera with 8 binary bit resolution, and the digital level is represented by a value of 0 to 255.
在用上述CCD相机拍摄的2维图像中，相当于光亮性颜料的反射光较强的部分的区域，由于闪亮感较强，因此亮度较高，在并非这样的部分的区域中，亮度较低。另外，即使相当于光亮性颜料的反射光较强的部分的区域，也因光亮性颜料的大小、形状、角度、材质等而亮度发生变化。即，可表示每个区域的亮度，本发明中根据各自区域中的亮度，可三维表示用CCD相机拍摄的2维图像的亮度分布。该亮度的三维分布图分为峰、谷及平坦部分，峰的高度或大小表示缘于光亮性颜料的光亮感的程度，表示峰越高则光亮感越显著，谷及平坦部分表示没有光亮感，表示主要由着色颜料或底材反射光。In the 2D image captured by the CCD camera, the area corresponding to the part where the reflected light of the bright pigment is strong has a higher brightness due to the strong shiny feeling, and the area not in such a part has a lower brightness. In addition, even the area corresponding to the part where the reflected light of the bright pigment is strong also changes in brightness due to the size, shape, angle, material, etc. of the bright pigment. That is, the brightness of each area can be represented, and the brightness distribution of the 2D image captured by the CCD camera can be represented in three dimensions according to the brightness in each area in the present invention. The three-dimensional distribution diagram of the brightness is divided into peaks, valleys and flat parts, and the height or size of the peaks represents the degree of the shiny feeling due to the bright pigment, indicating that the higher the peak, the more significant the shiny feeling, and the valleys and flat parts indicate that there is no shiny feeling, indicating that the light is mainly reflected by the coloring pigment or the base material.
关于用上述CCD相机拍摄的图像的解析，可用连接于CCD相机的图像解析装置来进行。作为在该图像解析装置中使用的图像解析软件，例如WinRoof(商品名：三谷商事公司制)等比较适合。The analysis of the image captured by the CCD camera can be performed using an image analysis device connected to the CCD camera. As image analysis software used in the image analysis device, WinRoof (trade name: manufactured by Mitani Shoji Co., Ltd.) is suitable.
在图像的解析中，分别定量性评价“闪亮感”(因从被测色表面中的光亮性颜料发生正反射的光而产生的不规则且细微的发光感)和“粒子感”(当难以发现闪亮感的照明条件下尽量观察被测色表面时，因光亮材质含有被测色表面中的光亮性颜料的定向、重叠而发生的不规则、无方向性的情形(随机模式)产生的感觉)，由于在目视观察中与观感评价的相关性也较高，测定时的缘于个人差的不均较小，因此比较适合。In the analysis of the image, the "shine" (the irregular and subtle luminescence caused by the light reflected from the bright pigment in the surface of the measured color) and the "graininess" (when the surface of the measured color is observed as much as possible under lighting conditions where the shine is difficult to detect, the irregular and non-directional situation (random pattern) caused by the orientation and overlap of the bright pigment in the bright material contained in the measured color surface) are quantitatively evaluated respectively. Since the correlation with the visual evaluation is also high in visual observation, the unevenness due to individual differences during measurement is small, so it is more suitable.
作为定量性测定闪亮感的适当方法，例如可例举下述的测定方法。将用CCD相机拍摄被光照射的被测色表面而形成的2维图像分割成多个区域，在该区域全部的跨度上得出该区域的各自的亮度的总计而得出合计值，该合计值除以全区域数而求出平均亮度x，将阈值α设定为该平均亮度x以上的值。优选阈值α通常为平均亮度x与y(y为24～40的数，优选28～36，更优选32)之和。As a suitable method for quantitatively measuring the sparkle, for example, the following measurement method can be cited. A two-dimensional image formed by photographing the surface of the color to be measured irradiated with light with a CCD camera is divided into a plurality of regions, and the total of the brightness of each region is obtained over the entire span of the region to obtain a total value, and the total value is divided by the number of all regions to obtain the average brightness x, and the threshold α is set to a value greater than the average brightness x. The preferred threshold α is usually the sum of the average brightness x and y (y is a number between 24 and 40, preferably between 28 and 36, and more preferably 32).
接下来，从上述区域各自的亮度减去阈值α的值，对该减算值为正值的该减算值进行合计，取得其总和即总体积V。另外，取得具有阈值α以上的亮度的区域的总数(通过用阈值α进行2值化而得到的具有上述阈值α以上的亮度的区域的总数)即总面积S。由于认为亮度峰值可近似于圆锥、角锥，因此通过总体积V除以总面积S的值的3倍，即通过下式Next, the threshold α is subtracted from the brightness of each of the above regions, and the subtraction values are summed up when the subtraction value is positive, and the total volume V is obtained. In addition, the total number of regions with brightness above the threshold α (the total number of regions with brightness above the threshold α obtained by binarizing the threshold α) is obtained, which is the total area S. Since the brightness peak is considered to be approximated to a cone or a pyramid, the total volume V is divided by 3 times the value of the total area S, that is, by the following formula:
PHavα＝3V/SPHavα＝3V/S
得到亮度峰值的平均高度PHavα。The average height PHavα of the brightness peak is obtained.
另外，设定上述平均亮度x以上且上述阈值α以下的阈值β。优选阈值β为阈值α以下，且通常为平均亮度x与z(z为16～32的数，优选20～28，更优选24)之和。In addition, a threshold β is set to be greater than the average brightness x and less than the threshold α. Preferably, the threshold β is less than the threshold α and is usually the sum of the average brightness x and z (z is a number of 16 to 32, preferably 20 to 28, and more preferably 24).
接下来，从上述区域各自的亮度减去阈值β的值，对该减算值为正值的该减算值进行合计，取得其总和即总体积W。另外，取得具有阈值β以上的亮度的区域的总数(通过用阈值β进行2值化而得到的上述阈值β以上的区域的总数)即总面积A。由于认为亮度峰值可近似于圆锥、角锥，因此通过总体积W除以总面积A的值的3倍，即通过下式Next, the threshold value β is subtracted from the brightness of each of the above regions, and the subtraction values are summed up when the subtraction value is positive, and the total volume W is obtained. In addition, the total number of regions with brightness above the threshold value β (the total number of regions above the threshold value β obtained by binarizing the threshold value β) is obtained, which is the total area A. Since it is believed that the brightness peak can be approximated to a cone or a pyramid, the total volume W is divided by 3 times the value of the total area A, that is, by the following formula
PHavβ＝3W/APHavβ＝3W/A
可得到阈值β的亮度峰值的平均高度PHavβ。The average height PHavβ of the brightness peak of the threshold β can be obtained.
另外，从阈值β的总面积A和表示阈值β以上的亮度的光学粒子的个数C，可求出光学粒子的平均粒子面积。本发明中，“光学粒子”表示“2维图像上亮度为阈值以上的独立的连续体”。假设上述光学粒子的形状为圆，用下述式式求出具有与平均粒子面积相同的面积的圆的直径D。In addition, the average particle area of the optical particles can be calculated from the total area A of the threshold value β and the number C of optical particles showing brightness above the threshold value β. In the present invention, "optical particles" means "independent continuum with brightness above the threshold value on a 2D image". Assuming that the shape of the above optical particles is a circle, the diameter D of a circle having the same area as the average particle area is calculated using the following formula.
数1Number 1
通过下述式Through the following formula
PSav＝D/PHavβPSav＝D/PHavβ
从上述PHavβ及D得出亮度峰值的平均变宽率PSav。The average broadening rate PSav of the brightness peak is obtained from the above PHavβ and D.
从如上所述地求出的亮度峰值的平均高度PHavα和如上所述地求出的亮度峰值的平均变宽率PSav，通过下述式From the average height PHavα of the brightness peaks obtained as described above and the average width variation rate PSav of the brightness peaks obtained as described above, the following formula is used:
BV＝PHavα+a·PSavBV＝PHavα+a·PSav
(式中，a为当PHavα小于25时300，当PHavα大于45时1050，当PHavα为25～45的数时用下述式(where a is 300 when PHavα is less than 25, 1050 when PHavα is greater than 45, and the following formula is used when PHavα is 25 to 45
a＝300+37.5×(PHavα-25)a＝300+37.5×(PHavα-25)
表示的值)The value represented by
可近似算出发光值BV。The luminous value BV can be approximately calculated.
在本发明的适当的方法中，通过如上所述地求出的发光值BV，可定量性测定光亮涂膜的“闪亮感”，即使在涂膜中的光亮材质的浓度差及亮度差较大时，发光值BV与目视观察的“闪亮感”的观感评价结果的相关性也比较高。In an appropriate method of the present invention, the "shining feeling" of the glossy coating can be quantitatively measured by the luminescence value BV calculated as described above. Even when the concentration difference and brightness difference of the glossy material in the coating are large, the correlation between the luminescence value BV and the visual evaluation result of the "shining feeling" observed visually is relatively high.
用MGR值表示粒子感。MGR值是微观观察时的质感即微观光亮感的尺度之一，是表示强光(相对于射入光从正反射附近观察多层涂膜)处的粒子感的参数。MGR值是如下得到的测定值，以入射角15度/受光角0度用CCD相机对多层涂膜的涂膜进行摄像，所得到的数码图像数据即对2维的亮度分布数据进行2维傅里叶变换处理，从所得到的功率谱图像只抽出对应于粒子感的空间频率区域，对所算出的计测参数进一步取出0至100的数值且以在与粒子感之间保持直线性关系方式进行变换。没有粒子感时成为0，最具粒子感时成为约100。The MGR value is used to represent the sense of granularity. The MGR value is one of the scales of the texture during microscopic observation, i.e., the sense of microscopic light, and is a parameter that represents the sense of granularity at strong light (observing the multilayer coating film from the vicinity of regular reflection relative to the incident light). The MGR value is a measured value obtained as follows: the coating film of the multilayer coating film is photographed with a CCD camera at an incident angle of 15 degrees/a light receiving angle of 0 degrees, and the obtained digital image data, i.e., the 2D brightness distribution data, is subjected to a 2D Fourier transform process. Only the spatial frequency region corresponding to the sense of granularity is extracted from the obtained power spectrum image, and the calculated measurement parameters are further taken out from 0 to 100 and transformed in a manner that maintains a linear relationship with the sense of granularity. It becomes 0 when there is no sense of granularity, and becomes about 100 when there is the most sense of granularity.
作为定量性测定“粒子感”的适当的方法可例举以下方法，如上所述地用CCD相机拍摄被光照射的光亮涂膜面而得到2维图像，从对该2维图像进行2维傅里叶变换而形成的空间频率光谱，得到将低空间频率成分的能量用积分及直流成分进行标准化而得到2维功率谱积分值，从该2维功率谱积分值定量性评价涂膜的粒子感。As an appropriate method for quantitatively measuring the "graininess", the following method can be cited. As described above, a CCD camera is used to capture a two-dimensional image of the bright coating surface irradiated with light. From the spatial frequency spectrum formed by performing a two-dimensional Fourier transform on the two-dimensional image, a two-dimensional power spectrum integral value is obtained by integrating the energy of the low spatial frequency components and normalizing the DC components. The particleiness of the coating is quantitatively evaluated from the two-dimensional power spectrum integral value.
从2维傅里叶变换后的空间频率光谱的图像抽出低空间频率成分，当测定对积分及直流成分进行标准化而得到的2维功率谱积分值时，将从空间频率光谱的图像抽出的低空间频率成分的抽出区域，做成表示析像度的线密度的下限值为0根/mm～上限值为2～13.4根/mm的范围内的任意数值的区域，优选做成0根/mm～4.4根/mm的区域，从与目视观察的“粒子感”的观感评价结果之间的相关性较高的观点考虑，比较适合。2维功率谱积分值越大则粒子感越大。The low spatial frequency component is extracted from the image of the spatial frequency spectrum after the 2D Fourier transform. When the integral value of the 2D power spectrum obtained by standardizing the integral and the DC component is measured, the extraction area of the low spatial frequency component extracted from the image of the spatial frequency spectrum is made into an area with a lower limit of 0 strands/mm and an upper limit of 2 to 13.4 strands/mm representing the linear density of the resolution, and preferably made into an area of 0 strands/mm to 4.4 strands/mm. Considering the high correlation with the visual evaluation result of the "graininess" observed by visual observation, it is more suitable. The larger the integral value of the 2D power spectrum, the greater the graininess.
可通过下式求出2维功率谱积分值(以下，有时简称为“IPSL”)。The two-dimensional power spectrum integral value (hereinafter, sometimes abbreviated as "IPSL") can be calculated by the following formula.
数2Number 2
(式中，ν为空间频率，θ为角度，P为功率谱，0～L为抽出的低空间频率区域，L表示抽出的频率的上限)(Where ν is the spatial frequency, θ is the angle, P is the power spectrum, 0 to L is the extracted low spatial frequency region, and L represents the upper limit of the extracted frequency)
另外，以所述发光值BV为基础，通过利用下述一次式In addition, based on the luminous value BV, by using the following formula
MBV＝(BV-50)/2MBV＝(BV-50)/2
计算的MBV值，还可以评价“闪亮感”。The calculated MBV value can also be used to evaluate the "shine".
MBV是没有闪亮感作为0且最具闪亮感作为约100的值，越具“闪亮感”则表示越大的数值。有时还将MBV值称为HB值(Hi-light Brilliant值)。MBV is a value where no brilliance is 0 and the most brilliance is about 100, and the more "brilliant" a piece is, the larger the value is. The MBV value is sometimes referred to as the HB value (Hi-light Brilliant value).
另外，还可以将所述2维功率谱积分值(IPSL)为基础，通过用下述一次式计算的MGR值评价“粒子感”。In addition, the "graininess" can also be evaluated by using the MGR value calculated by the following linear equation based on the two-dimensional power spectrum integrated value (IPSL).
当PSL值为0.32以上时，When the PSL value is above 0.32,
做成MGR＝[(IPSL×1000)-285]/2，MGR = [(IPSL × 1000) - 285] / 2,
当IPSL值处于0.15＜IPSL＜0.32的范围内时，When the IPSL value is in the range of 0.15＜IPSL＜0.32,
做成MGR＝[IPSL×(35/0.17)-(525/17)]/2MGR = [IPSL × (35/0.17) - (525/17)] / 2
当IPSL值为0.15以下时，做成MGR＝0。When the IPSL value is below 0.15, MGR=0.
上述MGR值是没有光亮材质的粒子感作为0且最具光亮材质的粒子感作为约100的值，越具“粒子感”则表示越大的数值。有时还将MGR值称为HG值(Hi-light Graininess值)。The MGR value is a value where no glossy texture is 0 and the glossiest texture is about 100, and the more "grainy" a texture is, the larger the value is. The MGR value is sometimes referred to as the HG value (Hi-light Graininess value).
另外，还可以通过指数化的数值(微观光亮感指数)，对根据上述MBV及MGR值综合表示微观光亮感的用下述式计算的微观光亮感进行评价。In addition, the microscopic brightness calculated by the following formula which comprehensively expresses the microscopic brightness based on the MBV and MGR values can also be evaluated by an indexed numerical value (microscopic brightness index).
微观光亮感指数＝(MGR+α·MBV)/(1+α)Microscopic brightness index = (MGR + α·MBV) / (1 + α)
通过对较多具有光亮感的被测色表面的研究，作为上述α值，如果优选1.80～1.40，更优选1.63，则可取得较好地与目视的微观光亮感一致的结果。微观光亮感指数是没有光亮感(闪亮感和粒子感均没有)作为0且最具光亮感(最具闪亮感和粒子感)作为约100的值Through the study of the measured color surfaces with a large number of glossy surfaces, it is found that if the α value is preferably 1.80 to 1.40, and more preferably 1.63, a result that is consistent with the visual microscopic glossiness can be obtained. The microscopic glossiness index is a value where no glossiness (no glossiness or graininess) is 0 and the most glossy (most glossy and grainy) is about 100.
配合组成数据Combination data
所述数据库中登记有所述组合物的配合组成数据。The database has registered therein the compounding composition data of the composition.
配合组成数据中包含所述组合物所含有的1种以上的色料、粘合剂、添加剂等的各配合成分及其各自的配合量所涉及的数据。另外，当作为组合物而使用市场上销售的产品时等，能够将商品名(品号)自身作为配合组成数据，另外还可以将各商品的配合量组成作为配合组成数据。例如，在用品号管理配合组成数据不明的市场上销售的产品或组合物时等有效。The matching composition data includes data related to each matching component and their respective matching amounts of one or more colorants, adhesives, additives, etc. contained in the composition. In addition, when a product sold on the market is used as a composition, the product name (article number) itself can be used as the matching composition data, and the matching amount composition of each product can also be used as the matching composition data. For example, it is effective when managing products or compositions sold on the market with unclear matching composition data by article number.
本发明中，关于组合物所含有的1种以上的色料、粘合剂、添加剂等的各成分，作为配合组成数据还可以登记其形状、化学特性等。作为形状，可例举色料等的形状(球状、鳞片状、纤维状等)、平均一次粒径、平均二次粒径、平均分散粒径、粒径分布、长宽比、厚度等。作为化学特性，可例举分子量、分子量分布、变色温度、反应性等。In the present invention, the shapes, chemical properties, etc. of the components of one or more colorants, binders, additives, etc. contained in the composition may be registered as the matching composition data. Examples of the shapes include the shapes of the colorants (spherical, flaky, fibrous, etc.), the average primary particle size, the average secondary particle size, the average dispersed particle size, the particle size distribution, the aspect ratio, the thickness, etc. Examples of the chemical properties include the molecular weight, the molecular weight distribution, the color change temperature, the reactivity, etc.
本发明中，当所述组合物中含有光反射性颜料或光干涉性颜料等光亮性颜料时，即使关于对光反射性颜料的含量、光干涉性颜料的含量、光亮性颜料的定向进行控制的定向控制剂的含量，也作为配合组成数据而登记在数据库中。In the present invention, when the composition contains a bright pigment such as a light reflective pigment or a light interference pigment, even the content of the light reflective pigment, the content of the light interference pigment, and the content of the orientation control agent for controlling the orientation of the bright pigment are registered in the database as compounding composition data.
而且，即使关于包含在所述组合物中的着色剂的各色相的含量、光反射性颜料的各色相的含量、光干涉性颜料的各色相的含量，也作为配合组成数据而登记在数据库中。Furthermore, the content of each hue of the colorant, the content of each hue of the light reflective pigment, and the content of each hue of the light interference pigment contained in the composition are also registered in the database as compounding composition data.
在此，关于各色相，可在1976年由国际照明委员会规定且JIS(日本工业标准)Z8729中也被采用的以L＊a＊b＊表色系为基础发明的L＊C＊h表色系中定义。Here, each hue can be defined in the L*C*h color system invented based on the L*a*b* color system, which was stipulated by the International Commission on Illumination in 1976 and adopted in JIS (Japanese Industrial Standards) Z8729.
例如，相对于以45度照射涂膜的光的正反射光，根据以45度受光时的分光反射率而计算的L＊C＊h表色系色度图中，将红色系的颜色定义为，在将a＊红色方向作为0度时，色相角度h处于-45度以上、小于45度的范围内的颜色。同样地，分别将橙色系的颜色定义为，在将a＊红色方向作为0度时，色相角度h处于45度以上、小于67.5度的范围内的颜色，将黄色系的颜色定义为，在将a＊红色方向作为0度时，色相角度h处于67.5度以上、小于135度的范围内的颜色，将绿色系的颜色定义为，在将a＊红色方向作为0度时，色相角度h处于135度以上、小于-135度的范围内的颜色，将青色系的颜色定义为，在将a＊红色方向作为0度时，色相角度h处于-135度以上、小于-45度的范围内的颜色。For example, in the L*C*h colorimetric chromaticity diagram calculated from the spectral reflectance when receiving light at 45 degrees with respect to regular reflected light irradiating the coating film at 45 degrees, red colors are defined as colors whose hue angle h is within the range of -45 degrees or more and less than 45 degrees when the a*red direction is taken as 0 degrees. Similarly, orange colors are defined as colors whose hue angle h is within the range of 45 degrees or more and less than 67.5 degrees when the a*red direction is taken as 0 degrees, yellow colors are defined as colors whose hue angle h is within the range of 67.5 degrees or more and less than 135 degrees when the a*red direction is taken as 0 degrees, green colors are defined as colors whose hue angle h is within the range of 135 degrees or more and less than -135 degrees when the a*red direction is taken as 0 degrees, and cyan colors are defined as colors whose hue angle h is within the range of -135 degrees or more and less than -45 degrees when the a*red direction is taken as 0 degrees.
涂装条件数据Coating condition data
所述数据库中还可以登记有涂装条件数据K。Painting condition data K may also be registered in the database.
涂装条件数据K是关于涂装的全部数据，例如可例举涂装中使用的涂装器具信息(涂装器具的种类、涂装器具的制造商、型号等)、涂装条件信息(涂装温度、涂装时湿度、干燥膜厚度、涂料固形成分、涂装距离、涂装速度等)、涂装者信息(姓名、涂装技能、涂装倾向、习惯等)、干燥条件信息(干燥温度、干燥湿度、干燥装置制造商、干燥装置型号)等。The coating condition data K is all data about coating, for example, it may include information on coating tools used in coating (type of coating tools, manufacturer and model of coating tools, etc.), coating condition information (coating temperature, humidity during coating, dry film thickness, paint solid content, coating distance, coating speed, etc.), painter information (name, coating skills, coating preferences, habits, etc.), drying condition information (drying temperature, drying humidity, drying device manufacturer, drying device model), etc.
计算机computer
由本发明中所使用的装置具备的计算机表示，超级计算机(supercomputer)、台式计算机、笔记本型计算机、便携式电脑等计算机及平板终端、智能手机等具有运算功能及信息处理功能的电子装置。计算机既可以设置在包括作业现场的任意场所，还可以由作业者等携带。The computer used in the present invention includes computers such as supercomputers, desktop computers, notebook computers, and portable computers, and electronic devices such as tablet terminals and smartphones that have computing and information processing functions. The computer can be set up in any place including the work site, and can also be carried by the operator.
计算机具备运算部及控制部，而且还可以具备输入输出部、通信部、记录部等。另外，本发明中，通过装配具备记录、运算、控制、输入输出功能等的色度计等，还可以与色度计等做成一体而实现。另外，还可以在计算机的记录部记录色彩数据及配合组成数据，在记录部设置数据库。The computer has a computing unit and a control unit, and may also have an input/output unit, a communication unit, a recording unit, etc. In addition, in the present invention, by assembling a colorimeter or the like having recording, computing, control, input/output functions, etc., it is also possible to realize the present invention by integrating the colorimeter or the like with the colorimeter or the like. In addition, the color data and the combination composition data may be recorded in the recording unit of the computer, and a database may be set in the recording unit.
在将数据库设置在计算机的外部时，计算机通过有线或无线连接于数据库。When the database is set outside the computer, the computer is connected to the database via wire or wirelessly.
计算机还可以通过有线或无线连接于用于测定颜色所涉及的数据及/或外观特性所涉及的数据的各种装置及自动调和机及其他运算装置、键盘、鼠标、读码器、触摸屏、图像识别装置等输入装置及监视器画面、印刷装置等输出装置等。The computer can also be connected via wired or wireless connection to various devices for measuring data related to color and/or data related to appearance characteristics, as well as automatic blenders and other computing devices, input devices such as keyboards, mice, barcode readers, touch screens, image recognition devices, and output devices such as monitor screens and printing devices.
根据需要，计算机中还可以安装(install)有用于运行本发明的方法或用于对本发明的系统进行控制并使其工作的应用程序软件(程序)，而进行必要的控制及工作。As required, application software (program) for running the method of the present invention or for controlling and operating the system of the present invention may be installed in the computer to perform necessary control and operations.
自动调和机Automatic blending machine
所述装置还可以具备自动调和机，其根据配合组成数据对各配合成分进行自动调和而调色。自动调和机可以通过有线或无线连接于所述计算机或数据库。The device may further include an automatic blender that automatically blends the blending components according to the blending composition data to adjust the color. The automatic blender may be connected to the computer or database via wired or wireless connection.
自动调和机至少具有：自动称量各色料等配合成分的重量或容量的电子天平；及将已被称量的各配合成分注入到调和机的注入器。The automatic blending machine at least has: an electronic balance for automatically weighing the weight or volume of each coloring material and other blending ingredients; and an injector for injecting the weighed blending ingredients into the blending machine.
通过使用所述自动调和机，能够自动进行高精度的称量，能够减少调和时的人为错误，能够迅速进行调和，能够容易调制任意量的调色结束的涂料。另外，通过调和作业的记录，能够容易进行生产管理。自动调和机既可以将调和所涉及的全部作业均自动化，另外还可以由作业者进行微调色等一部分。By using the automatic blending machine, high-precision weighing can be automatically performed, human errors during blending can be reduced, blending can be performed quickly, and any amount of color-mixed paint can be easily prepared. In addition, by recording the blending operation, production management can be easily performed. The automatic blending machine can automate all operations involved in blending, and can also allow the operator to perform part of the fine color adjustment.
基于计算机调色的涂料的制作方法(本发明的第1实施形态)Method for producing paint based on computer color matching (first embodiment of the present invention)
图6是当执行基于本发明的第1实施形态所涉及的计算机调色的涂料的制作方法时的流程图。并且，图6所示的流程只不过是本发明的一个实施方式。Fig. 6 is a flow chart of a method for producing a paint by computer coloring according to the first embodiment of the present invention. The flow chart shown in Fig. 6 is merely one embodiment of the present invention.
基于本发明的第1实施形态所涉及的计算机调色的涂料的制作方法是使用具备数据库和计算机的装置且包括下述S101～S111工序的方法，在该数据库中登记有1种以上的组合物C1～Cn(n为2以上的整数)的各自配合组成数据Y1～Yn及分别对应于各配合组成数据的色彩数据X1～Xn，在该计算机中利用登记在该数据库中的数据的配色计算逻辑发挥作用。The method for preparing a computer-colored paint according to the first embodiment of the present invention is a method using a device having a database and a computer and including the following steps S101 to S111, wherein respective combination composition data Y1 to Yn of one or more compositions C1 to Cn (n is an integer greater than 2) and color data X1 to Xn corresponding to each combination composition data are registered in the database, and a color matching calculation logic of the data registered in the database is utilized in the computer.
以下，对S101～S111工序进行详细说明。Hereinafter, steps S101 to S111 will be described in detail.
S101工序S101 process
S101工序是使用登记在所述数据库中的数据，向所述计算机输入学习用数据的工序。Step S101 is a step of inputting learning data into the computer using the data registered in the database.
本发明中，优选分别制作：使用1种以上的组合物且并不含有光亮性颜料的组合物的配合组成数据和色彩数据的学习用数据；及使用含有1种以上的色料、1种以上的光亮性颜料的组合物的配合组成数据和色彩数据的学习用数据，同时将这些分别输入。In the present invention, it is preferred to separately prepare: learning data for the combination composition data and color data of a composition using more than one composition and not containing a bright pigment; and learning data for the combination composition data and color data of a composition containing more than one colorant and more than one bright pigment, and input them separately at the same time.
本发明者根据光亮性颜料的有无，分别制作了学习用数据，通过分别输入这些数据，发现了计算机调色中的适合率格外提高。The inventors prepared learning data according to the presence or absence of bright pigments, and found that the adaptability rate in computer color matching was significantly improved by inputting these data.
本发明中，在将含有1种以上的色料和1种以上的光亮性颜料的组合物的配合组成数据作为学习用数据时，优选将选自光反射性颜料的含量、光干涉性颜料的含量、定向控制剂的含量及这些一个以上总计的1种以上的数据作为学习用数据来加以使用。In the present invention, when the compounding composition data of a composition containing one or more colorants and one or more bright pigments is used as learning data, it is preferred to use one or more data selected from the content of the light reflective pigment, the content of the light interference pigment, the content of the orientation control agent, and the total of one or more of these as learning data.
本发明中，在将含有1种以上的色料和1种以上的光亮性颜料的组合物的配合组成数据作为学习用数据时，优选将光亮性颜料的各色相的含量数据作为学习用数据。具体而言，优选将选自组合物中的光反射性颜料的各色相的含量、光干涉性颜料的各色相的含量及着色剂的各色相的含量的1种以上的数据作为学习用数据来加以使用。In the present invention, when the data of the compounding composition of a composition containing one or more colorants and one or more bright pigments is used as the learning data, the data of the content of each hue of the bright pigment is preferably used as the learning data. Specifically, it is preferred to use one or more data selected from the content of each hue of the light reflective pigment, the content of each hue of the light interference pigment, and the content of each hue of the colorant in the composition as the learning data.
关于光亮性颜料的颜色，可在1976年由国际照明委员会规定且JIS(日本工业标准)Z 8729中也被采用的以L＊a＊b＊表色系为基础发明的L＊C＊h表色系中定义。The color of the bright pigment can be defined in the L*C*h color system invented based on the L*a*b* color system, which was stipulated by the International Commission on Illumination in 1976 and adopted in JIS (Japanese Industrial Standards) Z 8729.
例如，相对于以45度照射涂膜的光的正反射光，根据以45度受光时的分光反射率而计算的L＊C＊h表色系色度图中，将红色系的颜色定义为，在将a＊红色方向作为0度时，色相角度h处于-45度以上、小于45度的范围内的颜色。同样地，分别将橙色系的颜色定义为，在将a＊红色方向作为0度时，色相角度h处于45度以上、小于67.5度的范围内的颜色，将黄色系的颜色定义为，在将a＊红色方向作为0度时，色相角度h处于67.5度以上、小于135度的范围内的颜色，将绿色系的颜色定义为，在将a＊红色方向作为0度时，色相角度h处于135度以上、小于-135度的范围内的颜色，将青色系的颜色定义为，在将a＊红色方向作为0度时，色相角度h处于-135度以上、小于-45度的范围内的颜色。For example, in the L*C*h colorimetric chromaticity diagram calculated from the spectral reflectance when receiving light at 45 degrees with respect to regular reflected light irradiating the coating film at 45 degrees, red colors are defined as colors whose hue angle h is within the range of -45 degrees or more and less than 45 degrees when the a*red direction is taken as 0 degrees. Similarly, orange colors are defined as colors whose hue angle h is within the range of 45 degrees or more and less than 67.5 degrees when the a*red direction is taken as 0 degrees, yellow colors are defined as colors whose hue angle h is within the range of 67.5 degrees or more and less than 135 degrees when the a*red direction is taken as 0 degrees, green colors are defined as colors whose hue angle h is within the range of 135 degrees or more and less than -135 degrees when the a*red direction is taken as 0 degrees, and cyan colors are defined as colors whose hue angle h is within the range of -135 degrees or more and less than -45 degrees when the a*red direction is taken as 0 degrees.
本发明中，当将1种以上的组合物的配合组成数据作为学习用数据时，优选将包含于组合物的色料的形状数据作为学习用数据。具体而言，优选将着色颜料及光亮性颜料等色料的形状(球状、鳞片状、纤维状等)、色料的平均一次粒径、平均二次粒径、平均分散粒径、粒径分布、长宽比、厚度等形状数据作为学习用数据而加以使用。In the present invention, when the data of the combination composition of more than one composition is used as the learning data, the shape data of the colorant contained in the composition is preferably used as the learning data. Specifically, the shape data of the colorant such as the shape (spherical, flaky, fibrous, etc.) of the colorant such as the coloring pigment and the bright pigment, the average primary particle size, the average secondary particle size, the average dispersed particle size, the particle size distribution, the aspect ratio, the thickness, etc. of the colorant are preferably used as the learning data.
向计算机的输入，可通过有线、无线或这些组合的通信手段或借由记录介质的手段来进行数据的发送。作为使用通信手段的输入，例如可例举LAN(局域网)、WAN(广域网络)、互联网、电话网等各种通信网络的1个以上的组合。作为通过借由记录介质的手段的输入，能够使用适当的读取手段读入磁记录介质、光学式记录介质、纸记录介质等记录介质的数据。The input to the computer can be transmitted by means of wired, wireless or a combination of these communication means or by means of a recording medium. As input using a communication means, for example, a combination of one or more of various communication networks such as a LAN (local area network), a WAN (wide area network), the Internet, and a telephone network can be cited. As input by means of a recording medium, data can be read into a recording medium such as a magnetic recording medium, an optical recording medium, and a paper recording medium using an appropriate reading means.
S102工序S102 process
S102是使用所述学习用数据进行机器学习，生成包含从色彩数据X推断配合组成数据Y的人工智能模型的至少1种的已学习的人工智能模型的工序。作为本发明中的人工智能模型，例如可由选自使用梯度提升的决策树、线形回归、逻辑回归、简单感知器、MLP、神经网络、支持向量机、随机森林、高斯过程、贝叶斯网络、k近邻法、其他机器学习中使用的模型的1种以上所构成。本发明中，优选由选自使用神经网络、梯度提升的决策树及高斯过程的1种以上所构成，尤其优选采用选自使用神经网络及梯度提升的决策树的1种以上的人工智能模型。S102 is a process of performing machine learning using the learning data to generate at least one learned artificial intelligence model including an artificial intelligence model that infers the matching composition data Y from the color data X. The artificial intelligence model in the present invention can be, for example, composed of one or more models selected from decision trees using gradient boosting, linear regression, logistic regression, simple perceptron, MLP, neural network, support vector machine, random forest, Gaussian process, Bayesian network, k-nearest neighbor method, and other models used in machine learning. In the present invention, it is preferably composed of one or more models selected from decision trees using neural networks, gradient boosting, and Gaussian processes, and it is particularly preferred to use one or more artificial intelligence models selected from decision trees using neural networks and gradient boosting.
本发明中，通过构成神经网络且使用S101工序中被输入的学习用数据使神经网络学习，能够生成包含从色彩数据X推断配合组成数据Y的人工智能模型的至少1种的已学习的人工智能模型。In the present invention, by constructing a neural network and using the learning data input in step S101 to make the neural network learn, it is possible to generate at least one learned artificial intelligence model including an artificial intelligence model that infers the combination composition data Y from the color data X.
本发明中，S102工序中生成的已学习的人工智能模型，在从色彩数据X推断配合组成数据Y的人工智能模型的至少1种的基础上，还可以包含从配合组成数据Y推断色彩数据X的人工智能模型的至少1种。在此，使用S101工序中被输入的学习用数据进行机器学习，而可生成从配合组成数据Y推断色彩数据X的人工智能模型的至少1种。即使在这样的情况下，也能够构成神经网络而使其学习。In the present invention, the learned artificial intelligence model generated in step S102 may include at least one artificial intelligence model that infers color data X from combination composition data Y, in addition to at least one artificial intelligence model that infers color data X from combination composition data Y. Here, machine learning is performed using the learning data input in step S101, and at least one artificial intelligence model that infers color data X from combination composition data Y can be generated. Even in such a case, a neural network can be constructed to learn.
使用S101工序中输入到计算机的学习用数据，实现人工智能模型(神经网络)的学习。学习用数据至少使用1种以上的组合物所涉及的色彩数据X和组成配合数据Y。作为神经网络的算法，可使用有教师学习方法之一即公知的误差逆传播算法。设定表示学习速度的参数即学习率(0～1之间的实数值)及学习中的输出值的误差的容许值即容许误差(0～1之间的实数值)，而使神经网络学习。由此，能够将配合组成数据Y所涉及的1种以上的特征量与基于该配合组成数据Y的涂料的涂膜的色彩数据X所涉及的1种以上的特征量关联起来。使用已被学习的网络，通过前馈计算能够预测满足色彩数据的组成配合数据及基于组成配合数据的色彩数据。已被学习的网络，并不进行费用或时间等工时所涉及的实验性确认而预测这些。The learning data input into the computer in the S101 process is used to realize the learning of the artificial intelligence model (neural network). The learning data uses at least one or more color data X and composition data Y related to the composition. As the algorithm of the neural network, one of the teacher learning methods, namely the well-known error back propagation algorithm, can be used. The parameters representing the learning speed, namely the learning rate (real value between 0 and 1) and the allowable value of the error of the output value during learning, namely the allowable error (real value between 0 and 1), are set to make the neural network learn. In this way, it is possible to associate one or more feature quantities related to the combination composition data Y with one or more feature quantities related to the color data X of the coating film of the paint based on the combination composition data Y. Using the learned network, the composition data that satisfies the color data and the color data based on the composition data can be predicted by feedforward calculation. The learned network predicts these without experimental confirmation involving cost or time and other labor hours.
本发明中，优选生成所述已学习的人工智能模型的工序包括：(i)作为学习用数据而使用并不含有光亮性颜料的组合物的1种以上所涉及的各配合组成数据Y和各色彩数据X，使人工智能模型学习的工序；及(ii)作为学习用数据而使用含有光亮性颜料的组合物的1种以上所涉及的各配合组成数据Y和各色彩数据X，使人工智能模型学习的工序。此时，优选工序(ii)包括，将选自组合物中的光反射性颜料的含量、光干涉性颜料的含量、定向控制剂的含量及这些一个以上总计的1种以上的数据作为学习用数据来加以使用，而使人工智能模型学习的工序。In the present invention, the process of generating the learned artificial intelligence model preferably includes: (i) using as learning data the combination composition data Y and color data X of one or more compositions that do not contain bright pigments to make the artificial intelligence model learn; and (ii) using as learning data the combination composition data Y and color data X of one or more compositions that contain bright pigments to make the artificial intelligence model learn. In this case, preferably, step (ii) includes the process of using as learning data the data of one or more selected from the content of light reflective pigments, the content of light interference pigments, the content of orientation control agents, and the total of one or more of these in the composition to make the artificial intelligence model learn.
在使人工智能模型学习的工序中，并不特意限制所述工序(i)与工序(ii)的顺序，而是既可以在工序(i)之后执行工序(ii)，还可以在工序(ii)之后执行工序(i)。本发明中，从组合物是否含有光亮性颜料的观点考虑，分别制作学习用数据，通过使其学习，能够生成以更高精度可进行预测的人工智能模型。In the process of making the artificial intelligence model learn, the order of the steps (i) and (ii) is not particularly limited, and the step (ii) can be performed after the step (i), or the step (i) can be performed after the step (ii). In the present invention, learning data is prepared from the perspective of whether the composition contains a bright pigment, and by making it learn, an artificial intelligence model that can make predictions with higher accuracy can be generated.
本发明中，优选生成所述已学习的人工智能模型的工序包括，将选自组合物中的光反射性颜料的含量、光干涉性颜料的含量、不定形二氧化硅等定向控制剂的含量及这些1个以上总计的1种以上的数据作为学习用数据而加以使用，而使人工智能模型学习的工序。由此，能够取得尤其高精度地可适合于含有光亮性颜料的被测色表面的色彩数据的配合组成数据。In the present invention, the step of generating the learned artificial intelligence model preferably includes a step of using one or more data selected from the content of light-reflecting pigments, the content of light-interference pigments, the content of orientation control agents such as amorphous silica, and the total of one or more of these in the composition as learning data to make the artificial intelligence model learn. In this way, it is possible to obtain the color data of the color surface to be measured containing bright pigments, which is particularly suitable with high accuracy.
本发明中，优选生成所述已学习的人工智能模型的工序包括，将选自组合物中的光反射性颜料的各色相的含量、光干涉性颜料的各色相的含量及着色剂的各色相的含量的1种以上的数据作为学习用数据而加以使用，而使人工智能模型学习的工序。由此，能够取得尤其更加高精度地可适合于含有光亮性颜料的被测色表面的色彩数据的配合组成数据。In the present invention, the step of generating the learned artificial intelligence model preferably includes a step of using one or more data selected from the content of each hue of the light reflective pigment, the content of each hue of the light interference pigment, and the content of each hue of the colorant in the composition as learning data to make the artificial intelligence model learn. In this way, it is possible to obtain the combination composition data that is particularly suitable for the color data of the measured color surface containing the bright pigment with higher accuracy.
本发明中，优选生成所述已学习的人工智能模型的工序包括，将包含于组合物的色料的形状数据作为学习用数据而加以使用，而使人工智能模型学习的工序。由此，能够使被测色表面的色彩数据中的起因于粒子的感觉更加高精度地适合。In the present invention, the step of generating the learned artificial intelligence model preferably includes a step of using the shape data of the colorant contained in the composition as learning data to make the artificial intelligence model learn. Thus, the feeling caused by particles in the color data of the color surface to be measured can be more accurately adapted.
本发明中，色料并不仅是无机着色颜料或有机着色颜料等通常的着色剂，而是还包括粒子状或薄片状(鳞片状)的玻璃、金属、二氧化硅、氧化铝等的光亮性颜料或薄片状的具有干涉性的玻璃(例如，二氧化硅覆盖玻璃薄片等)、二氧化硅、氧化铝等光干涉性颜料。另外，形状数据包括球状、薄片状、纤维状等形状外观及粒径、粒度分布、厚度、长宽比、纤维长、纤维径等数据。In the present invention, the colorant is not only a common colorant such as an inorganic coloring pigment or an organic coloring pigment, but also includes a bright pigment such as glass, metal, silicon dioxide, aluminum oxide, etc. in a particle or flaky (scaly) form, or a light interference pigment such as glass with interference properties in a flaky form (e.g., glass flakes covered with silicon dioxide, etc.), silicon dioxide, aluminum oxide, etc. In addition, the shape data includes the shape appearance such as spherical, flaky, fibrous, and the data of particle size, particle size distribution, thickness, aspect ratio, fiber length, fiber diameter, etc.
S103工序S103 process
S103工序是取得配合组成Yp为未知的作为目标的色彩的色彩数据Xp(以下，有时称为“目标色彩数据Xp”)的工序。Step S103 is a step of acquiring color data Xp (hereinafter sometimes referred to as “target color data Xp”) of a target color whose blending composition Yp is unknown.
作为目标色彩数据Xp，可例举关于涂装物、成形品、自然结构物等所具有的全部色彩的色彩数据。尤其，优选作为涂装物的色彩数据。The target color data Xp may be color data on all colors of painted objects, molded products, natural structures, etc. In particular, color data on painted objects is preferred.
即使将目前为止难以进行计算机调色的含有光亮性颜料的涂膜的色彩数据作为目标色彩数据Xp，本发明也能够高精度地进行调色。因此，优选S103工序中的目标色彩数据Xp为含有光亮性颜料的涂膜的色彩数据。当然，S103工序中的目标色彩数据Xp还可以为并不含有光亮性颜料的涂膜的色彩数据。Even if the color data of a coating film containing bright pigments, which has been difficult to color by computer so far, is used as the target color data Xp, the present invention can also color-match with high precision. Therefore, it is preferred that the target color data Xp in step S103 is the color data of a coating film containing bright pigments. Of course, the target color data Xp in step S103 can also be the color data of a coating film that does not contain bright pigments.
构成目标色彩数据Xp的要素，可以与构成登记在数据库中的色彩数据的要素相同。例如，可以为通过计量仪器测定的色彩数据或从其算出的色彩数据。The elements constituting the target color data Xp may be the same as the elements constituting the color data registered in the database, for example, the color data measured by a measuring instrument or the color data calculated therefrom.
作为用于取得色彩数据的计量仪器，只要是可测定光亮涂膜(金属涂膜、珍珠色涂膜等)、纯色涂膜等的色彩而取得色彩数据的计量仪器，则并不特意限制采用测定原理、测定值的色彩数据的算出方法等，而是可使用现有公知的计量仪器。例如，可使用具备照射被测色表面的光源的单角度分光光度计、多角度分光光度计、色彩计、色差计、变角分光光度计等色度计及摄像装置、微观光亮感测定仪等测定仪器及色样本卡等计量仪器的1个以上。另外，可任意使用对从这些计量仪器得到的各种色彩数据进行处理的数据处理装置。As a measuring instrument for obtaining color data, as long as it can measure the color of a bright coating (metallic coating, pearl coating, etc.), a solid color coating, etc. and obtain color data, there is no particular limitation on the measurement principle adopted, the method of calculating the color data of the measured value, etc., and any existing known measuring instrument can be used. For example, a colorimeter such as a single-angle spectrophotometer, a multi-angle spectrophotometer, a colorimeter, a colorimeter, a variable-angle spectrophotometer, and a camera device, a measuring instrument such as a microscopic brightness meter, and a color sample card can be used. In addition, a data processing device that processes various color data obtained from these measuring instruments can be used arbitrarily.
作业者使用各种计量仪器，直接测定被测色物来可得到目标色彩数据Xp。另外，还可以由各种计量仪器根据程序等自动取得。而且，还可以根据这些测色数据进行算出。The operator can obtain the target color data Xp by directly measuring the color object using various measuring instruments. In addition, it can also be automatically obtained by various measuring instruments according to programs, etc. Moreover, it can also be calculated based on these color measurement data.
本发明中，优选使用多角度分光光度计对被测色表面进行测定，而取得目标色彩数据Xp。In the present invention, it is preferred to use a multi-angle spectrophotometer to measure the color surface to be measured and obtain the target color data Xp.
另外，当目标色彩数据Xp并不是直接测定被测色物而得到的数据时，能够将从被测色物的商品名等得到的色彩数据作为目标色彩数据Xp而加以使用。例如，当目标色彩数据Xp为有关汽车的色彩数据时，可根据汽车的商品名、型号、年式、制造号等得到的涂料数据而设定目标色彩数据Xp。In addition, when the target color data Xp is not data obtained by directly measuring the color object to be measured, color data obtained from the product name of the color object to be measured, etc. can be used as the target color data Xp. For example, when the target color data Xp is color data related to a car, the target color data Xp can be set based on paint data obtained from the product name, model, year, manufacturing number, etc. of the car.
S104工序S104 process
S104工序是向所述计算机输入所述目标色彩数据Xp的工序。Step S104 is a step of inputting the target color data Xp into the computer.
向计算机的输入，可通过有线、无线或这些组合的通信手段或借由记录介质的手段，从测定及/或算出色彩数据Xp的各种装置发送数据并使计算机接收。作为使用通信手段的输入，例如可例举LAN(局域网)、WAN(广域网络)、互联网、电话网等各种通信网络的1个以上的组合。作为通过借由记录介质的手段的输入，能够使用适当的读取手段读入将磁记录介质、光学式记录介质、纸记录介质等记录介质的数据。The input to the computer can be transmitted from various devices that measure and/or calculate the color data Xp and received by the computer through wired, wireless or a combination of these communication means or through a recording medium. As input using a communication means, for example, a combination of one or more communication networks such as a LAN (local area network), a WAN (wide area network), the Internet, and a telephone network can be cited. As input through a recording medium, data can be read from a recording medium such as a magnetic recording medium, an optical recording medium, or a paper recording medium using an appropriate reading means.
另外，可使用键盘、鼠标、读码器、触摸屏、声音输入装置、图像识别装置等连接于所述计算机或由计算机所具备输入手段来输入。In addition, input can be performed using a keyboard, a mouse, a code reader, a touch screen, a voice input device, an image recognition device, or the like connected to the computer or input by an input means provided by the computer.
S105工序S105 process
S105工序是使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，将从色彩数据Xp预测的预测配合组成数据Ya1，作为将所述组合物C1～Cn的1种以上作为成分的配合组成数据而取得的工序。Step S105 is a step of obtaining predicted combination composition data Ya1 predicted from color data Xp as combination composition data having one or more of the compositions C1 to Cn as components, using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model.
S105工序中使用的所述已学习的人工智能模型是在S102工序中生成的已学习的人工智能模型，是从色彩数据X推断配合组成数据Y的人工智能模型的至少1种。The learned artificial intelligence model used in step S105 is the learned artificial intelligence model generated in step S102, and is at least one artificial intelligence model that infers the matching composition data Y from the color data X.
作为使用已被学习的人工智能模型将从色彩数据Xp预测的预测配合组成数据Ya1作为所述组合物C1～Cn的1种以上的成分的配合组成数据而得到的方法，向已被学习的人工智能模型的神经网络中的输入层的各单元，输入色彩数据Xp的特征量即可。输入到输入层的色彩数据Xp，在各节点及各层之间一边被加权一边被发送，从输出层的各单元作为预测配合组成数据Ya1而被输出。As a method for obtaining the predicted combination composition data Ya1 predicted from the color data Xp as the combination composition data of one or more components of the composition C1 to Cn using the learned artificial intelligence model, the feature amount of the color data Xp is input to each unit of the input layer in the neural network of the learned artificial intelligence model. The color data Xp input to the input layer is transmitted while being weighted between each node and each layer, and is output from each unit of the output layer as the predicted combination composition data Ya1.
S105工序可包括，采用多标签分类能够将从色彩数据Xp预测的预测配合组成数据Ya1作为以所述组合物C1～Cn的1种以上作为成分的配合组成数据而得到的工序。还可以将这样的预测配合组成数据Ya1作为所述组合物C1～Cn的2种以上的组成而求出(例如，C1的量、C2的量···Cn的量等，作为基于组合物单位的组成而求出)，另外，还可以作为构成所述组合物C1～Cn的各成分(颜料等)的各自的量而求出(例如，红色颜料A的量、红色颜料B的量等，作为基于各构成成分的组成而求出)。The step S105 may include a step of using multi-label classification to obtain the predicted combination composition data Ya1 predicted from the color data Xp as combination composition data with one or more of the compositions C1 to Cn as components. Such predicted combination composition data Ya1 may also be obtained as two or more compositions of the compositions C1 to Cn (for example, the amount of C1, the amount of C2, the amount of Cn, etc., as a composition based on a composition unit), and may also be obtained as the respective amounts of each component (pigment, etc.) constituting the compositions C1 to Cn (for example, the amount of red pigment A, the amount of red pigment B, etc., as a composition based on each constituent component).
在此，多标签分类被设定成，对于特定的对象物，同时存在2个以上的解答(标签)，或者同时可分配到2个以上的解答(类别)。Here, multi-label classification is set so that for a specific object, there are two or more solutions (labels) at the same time, or two or more solutions (categories) can be assigned at the same time.
成为特定的色彩数据的配合组成，通常并不只存在1个，而是存在多个。因此，通过采用多标签分类，能够更加高效地得到从色彩数据Xp预测的预测配合组成数据Ya1。例如，通过采用多标签分类，当求出满足绿色系的色彩数据Xg的配合组成数据时，作为绿色系组合物Cg1和Cg2的组合物，能够将配合组成数据作为解，作为进一步的解，作为黄色系组合物Cy1和青色系组合物Cb1的组合物，能够将配合组成数据作为解。There is usually not just one but multiple matching compositions that become specific color data. Therefore, by adopting multi-label classification, the predicted matching composition data Ya1 predicted from the color data Xp can be obtained more efficiently. For example, by adopting multi-label classification, when the matching composition data of the color data Xg that satisfies the green system is obtained, the matching composition data can be used as a solution as a combination of green system compositions Cg1 and Cg2, and as a further solution, the matching composition data can be used as a solution as a combination of yellow system composition Cy1 and cyan system composition Cb1.
本发明中，通过采用多标签分类，也能够以概率表示预测配合组成数据提供的各成分的存在量，由此能够高概率地预测在预测配合组成数据中的各成分组成(基于组合物单位的组成及基于各构成成分的组成这两者)。In the present invention, by adopting multi-label classification, the presence amount of each component provided by the predicted combination composition data can also be expressed in terms of probability, thereby being able to predict the composition of each component in the predicted combination composition data (both the composition based on the combination unit and the composition based on each constituent component) with high probability.
作为使用S105工序中使用的人工智能模型以外的预测式的方法，例如可例举作为计算机色彩校正(CCM)而已周知的方法，即基于使用计算机的配色计算逻辑的算出或基于数理最优化的算出。As a prediction method other than the artificial intelligence model used in the S105 process, for example, a method known as computer color correction (CCM), that is, calculation based on color matching calculation logic using a computer or calculation based on mathematical optimization can be cited.
S105工序还可以作为相当于计算机选色(CCS)的工序。例如，能够在登记于数据库的多个色彩数据中，检索近似于目标色彩数据Xp的色彩数据，在作为检索色彩数据Xn1而取得之后，将对应于检索色彩数据Xn1的配合组成数据作为预测配合组成数据。The step S105 can also be used as a process equivalent to computer color selection (CCS). For example, color data similar to the target color data Xp can be retrieved from a plurality of color data registered in the database, and after being obtained as the retrieved color data Xn1, the matching composition data corresponding to the retrieved color data Xn1 can be used as the predicted matching composition data.
在此，登记于数据库的色彩数据例如为公知的色样本账的色彩数据或过去制作的涂板的色彩数据等，均与对应于色彩数据的配合组成数据有关联。从而，通过取得检索色彩数据Xn1，将对应的配合组成数据作为预测配合组成数据而能够容易得到。Here, the color data registered in the database, such as the color data of a known color sample book or the color data of a paint plate made in the past, are all associated with the matching composition data corresponding to the color data. Therefore, by obtaining the search color data Xn1, the corresponding matching composition data can be easily obtained as the predicted matching composition data.
分别对构成色彩数据的要素的1个以上(例如，L＊a＊b＊表色系中的各值等)，与构成目标色彩数据Xp的对应的要素进行对比，能够通过检索值的差分、一致度、误差率等处于一定的范围内的要素来取得检索色彩数据Xn1。所述一定的范围既可以由作业者参考经验等来设定，另外还可以通过计算机来设定。By comparing one or more elements constituting the color data (e.g., each value in the L*a*b* colorimetric system) with the corresponding elements constituting the target color data Xp, the search color data Xn1 can be obtained by searching for elements whose difference, consistency, error rate, etc. of the search value are within a certain range. The certain range can be set by the operator based on experience or by a computer.
S105工序中，当比较目标色彩数据Xp与所述检索色彩数据Xn1而判定合格与否时，可着眼于构成检索色彩数据Xn1的要素的1个以上与构成目标色彩数据Xp的要素的1个以上，对分别对应的各构成要素进行比较来进行。当判定合格与否时，例如还可以对各构成要素中的差分、一致度、误差率等设置阈值，参考这些由仪器或作业者判定合格与否。此时，还可以反映熟练作业者的观点等，而在各构成要素之间进行加权。In step S105, when comparing the target color data Xp with the search color data Xn1 to determine whether it is qualified or not, it is possible to focus on one or more elements constituting the search color data Xn1 and one or more elements constituting the target color data Xp, and compare the corresponding constituent elements. When determining whether it is qualified or not, for example, a threshold value can be set for the difference, consistency, error rate, etc. in each constituent element, and the instrument or operator can refer to these to determine whether it is qualified or not. At this time, it is also possible to reflect the opinions of skilled operators and weight the constituent elements.
能够将在所述S105工序中得到的预测配合组成数据Ya1作为所述组合物C1～Cn的1种以上的组合物的数据。另外，还可以作为树脂、色料、溶剂等各构成要素的配合量比的数据。优选考虑实际配合时的作业性等，而作为所述组合物C1～Cn的1种以上的组合物的数据。The predicted compounding composition data Ya1 obtained in the step S105 can be used as data of one or more compositions of the compositions C1 to Cn. In addition, it can be used as data of compounding ratios of components such as resins, colorants, and solvents. It is preferred to use the data of one or more compositions of the compositions C1 to Cn in consideration of workability during actual compounding.
在S105工序中得到的预测配合组成数据Ya1中，虽然并不特意限制所述组合物C1～Cn的种类，但是考虑实际配合时的作业性等，而可做成15种以下，优选做成12种以下，更优选做成10种以下。此时，作为含有金属颜料的组合物，做成5种以下，优选做成3种以下，作为含有珠光颜料的组合物，做成5种以下，优选做成3种以下。In the predicted compounding composition data Ya1 obtained in step S105, the types of the compositions C1 to Cn are not particularly limited, but in consideration of workability during actual compounding, the types may be 15 or less, preferably 12 or less, and more preferably 10 or less. In this case, the types of compositions containing metallic pigments may be 5 or less, preferably 3 or less, and the types of compositions containing pearlescent pigments may be 5 or less, preferably 3 or less.
优选本发明中的S105工序为，使用所述已学习的人工智能模型即从色彩数据X推断配合组成数据Y的人工智能模型的至少1种，将从色彩数据Xp预测的预测配合组成数据Ya1，作为将所述组合物C1～Cn的1种以上作为成分的配合组成数据而得到的工序。Preferably, the S105 process in the present invention is a process in which at least one of the learned artificial intelligence models, i.e., the artificial intelligence models for inferring the combination composition data Y from the color data X, is used to obtain the predicted combination composition data Ya1 predicted from the color data Xp as the combination composition data with one or more of the compositions C1 to Cn as components.
S106工序S106 process
S106工序是使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，取得从预测配合组成数据Ya1预测的预测色彩数据Xa1，同时与所述色彩数据Xp进行比较而判定合格与否的工序。Process S106 is a process of using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model to obtain the predicted color data Xa1 predicted from the predicted combination composition data Ya1, and compare it with the color data Xp to determine whether it is qualified or not.
作为在取得从预测配合组成数据Ya1预测的预测色彩数据Xa1的方法中所使用的所述已学习的人工智能模型，可例举在S102工序中生成的已学习的人工智能模型且从配合组成数据Y推断色彩数据X的人工智能模型的至少1种。As the learned artificial intelligence model used in the method of obtaining the predicted color data Xa1 predicted from the predicted combination composition data Ya1, there can be cited at least one of the learned artificial intelligence models generated in step S102 and the artificial intelligence model that infers the color data X from the combination composition data Y.
作为取得已被学习的人工智能模型的预测色彩数据Xa1的方法，向已被学习的人工智能模型的神经网络中的输入层的各单元，输入预测配合组成数据Ya1的特征量即可。输入到输入层的预测配合组成数据Ya1，在各节点及各层之间一边被加权一边被发送，从输出层的各单元作为预测色彩数据Xa1而被输出。As a method for obtaining the predicted color data Xa1 of the learned artificial intelligence model, the feature value of the predicted combination composition data Ya1 can be input to each unit of the input layer in the neural network of the learned artificial intelligence model. The predicted combination composition data Ya1 input to the input layer is transmitted while being weighted between each node and each layer, and is output from each unit of the output layer as the predicted color data Xa1.
作为在取得从预测配合组成数据Ya1预测的预测色彩数据Xa1的方法中所使用的使用人工智能模型以外的预测式的方法，例如可例举作为计算机色彩校正(CCM)而已周知的方法，即基于使用计算机的配色计算逻辑的算出或基于数理最优化的算出。As a method of using a prediction formula other than an artificial intelligence model used in the method of obtaining the predicted color data Xa1 predicted from the predicted combination composition data Ya1, for example, a method known as computer color correction (CCM), that is, calculation based on color matching calculation logic using a computer or calculation based on mathematical optimization can be cited.
基于使用计算机的配色计算逻辑的算出，例如基于登记在所述数据库中的各种色彩数据以及与其对应的组成配合数据，对目标色彩数据Xp与所述各种色彩数据进行比较而以差分、一致度等处于一定的范围的方式进行计算，由此将认为最合理的一个以上的配合组成决定为预测配合组成数据Ya1。利用构成计算逻辑的各种函数，通过较小的反复工序可修正任意的配合组成或近似配合组成。此时，以有规则的形态生成理论指令，能够补助计算速度或调整算法的精度。Based on the calculation of the color matching calculation logic using a computer, for example, based on the various color data registered in the database and the corresponding composition matching data, the target color data Xp is compared with the various color data and the calculation is performed in a manner such that the difference, consistency, etc. are within a certain range, thereby determining one or more matching compositions that are considered to be the most reasonable as the predicted matching composition data Ya1. Using various functions that constitute the calculation logic, any matching composition or approximate matching composition can be corrected through a small repetitive process. At this time, generating theoretical instructions in a regular form can supplement the calculation speed or adjust the accuracy of the algorithm.
预测配合组成数据Ya1例如参照记录在数据库中的色彩数据，对每一个构成通过近似配合组成得到的色彩数据的坐标轴，检索具有在减小误差的方向上发挥作用的特性信息的成分，由此可得到通过数理最优化的算出。例如，L＊a＊b＊表色系中，当将误差作为ΔL＊＝L＊2-L＊1、Δa＊＝a＊2-a＊1、Δb＊＝b＊2-b＊1时，当L＊轴上的误差ΔL＊为正值时，检索具有在减小L＊2值的方向上发挥作用的特性信息的成分，当L＊轴上的误差ΔL＊为负值时，检索具有增加L＊2值的特性信息的成分。同样地，分别当a＊轴上误差Δa＊为正值时，检索具有减小a＊2值的特性信息(绿)的成分，当a＊轴上的误差Δa＊为负值时，检索具有增加a＊2值的特性信息(红)的成分，当b＊轴上的误差Δb＊为正值时，检索具有减小b＊2值的特性信息(青)的成分，当b＊轴上的误差Δb＊为负值时，检索具有增加b＊2值的特性信息(黄)的成分。由此，近似配合组成中，对每一个构成色空间的表色系的坐标轴，通过追加在减小误差的方向上发挥作用即用于赋予规定的特性信息的成分，能够取得接近作为目标的色彩的预测配合组成数据Ya1。The predicted combination composition data Ya1 refers to the color data recorded in the database, for example, and for each coordinate axis constituting the color data obtained by the approximate combination composition, a component having characteristic information that acts in the direction of reducing the error is retrieved, thereby obtaining a calculation by mathematical optimization. For example, in the L*a*b* color system, when the error is ΔL*=L*2 -L*1 , Δa*=a*2 -a*1 , Δb*=b*2 -b*1 , when the error ΔL* on the L* axis is a positive value, a component having characteristic information that acts in the direction of reducing the L*2 value is retrieved, and when the error ΔL* on the L* axis is a negative value, a component having characteristic information that increases the L*2 value is retrieved. Similarly, when the error Δa* on the a* axis is a positive value, a component having characteristic information (green) that reduces the a*2 value is retrieved, when the error Δa* on the a* axis is a negative value, a component having characteristic information (red) that increases the a*2 value is retrieved, when the error Δb* on the b* axis is a positive value, a component having characteristic information (cyan) that reduces the b*2 value is retrieved, and when the error Δb* on the b* axis is a negative value, a component having characteristic information (yellow) that increases the b*2 value is retrieved. Thus, in the approximate matching composition, by adding a component that acts in the direction of reducing the error, i.e., for providing a predetermined characteristic information, to each coordinate axis of the color system constituting the color space, it is possible to obtain predicted matching composition data Ya1 close to the target color.
假设，当并未检索到具有在减小误差的方向上发挥作用的特性信息的成分时，根据目标色彩数据Xp，能够在得到更适合的新的近似配合组成之后取得候补配合组成数据。Assuming that a component having characteristic information that works in a direction of reducing the error is not found, candidate blend composition data can be acquired after a more suitable new approximate blend composition is obtained based on the target color data Xp.
另外，还可以对通过CCM得到的配合组成数据进行作业者的修正(例如，公知的色样本账的色彩数据、过去制作的涂板的色彩数据、参考自身经验等而进行的作业者的修正)、所述计算机的修正、人工智能模型的修正等而得到预测配合组成数据Ya1。In addition, the combination composition data obtained through CCM can also be corrected by the operator (for example, color data of a known color sample account, color data of a coating plate made in the past, corrections made by the operator based on his own experience, etc.), corrections by the computer, corrections by the artificial intelligence model, etc. to obtain the predicted combination composition data Ya1.
而且，在汽车修理工场等的作业现场中，当可使用的色料或组合物被限制时等，还可以只根据作业现场中可使用的色料或组合物而取得预测配合组成数据Ya1。Furthermore, when the colorants or compositions that can be used are limited at a work site such as an automobile repair shop, the predicted blending composition data Ya1 can be acquired based on only the colorants or compositions that can be used at the work site.
可通过显示手段或印刷手段等输出预测配合组成数据Ya1。另外，还可以并不输出，而是从计算机发送至实施下一个工序的仪器等。The predicted compound composition data Ya1 may be outputted by display means or printing means, etc. Alternatively, the data may be transmitted from the computer to a device or the like that performs the next step instead of being outputted.
作为使用已被学习的人工智能模型以外的预测式取得预测色彩数据Xa1的方法，可使用利用CCM的调色领域中的公知的各种预测式。作为这样的预测式例如可例举采用基于Kubelka-Munk光学浓度式和Duncan混色理论式的2定数法的预测式的方法，采用模糊推论的方法，以及其他通过计算机对色彩数据或配合组成数据进行比较而将各自的匹配程度指数化的方法等。As a method of obtaining the predicted color data Xa1 using a prediction formula other than the learned artificial intelligence model, various prediction formulas known in the field of color matching using CCM can be used. Examples of such prediction formulas include a method using a prediction formula based on the Kubelka-Munk optical density formula and the 2 constants method of the Duncan color mixing theory formula, a method using fuzzy inference, and other methods that compare color data or matching composition data by computer and index the degree of matching.
采用Kubelka-Munk光学浓度式和Duncan混色理论式的方法如下。求出包含于1种以上的组合物的各色料的各自光散射系数、光吸收系数和各色料的配合比率，从Kubelka-Munk光学浓度式算出“混色后的光吸收系数/混色后的光散射系数”，使用该值通过Duncan混色理论式可求出分光反射率。另外，可以从作为目标的色彩的分光反射率算出“光吸收系数/光散射系数”，求出与此对齐颜色所需的色料或组合物等各原色涂料的配合比。通过对可见光谱的各波长进行该计算，能够决定用于做成作为目标的色彩的颜料配合比。The method using the Kubelka-Munk optical concentration formula and the Duncan color mixing theory formula is as follows. The light scattering coefficient, light absorption coefficient and mixing ratio of each colorant contained in one or more compositions are calculated, and the "light absorption coefficient after color mixing/light scattering coefficient after color mixing" is calculated from the Kubelka-Munk optical concentration formula. The spectral reflectance can be calculated using the Duncan color mixing theory formula using this value. In addition, the "light absorption coefficient/light scattering coefficient" can be calculated from the spectral reflectance of the target color to find the mixing ratio of each primary color coating such as the colorant or composition required to match the color. By performing this calculation for each wavelength of the visible spectrum, the pigment mixing ratio used to make the target color can be determined.
在此，Kubelka-Munk光学浓度式如同以下所示。Here, the Kubelka-Munk optical density formula is as shown below.
数3Number 3
(K/S)λ：波长λ的Kubelka-Munk光学浓度函数(K/S)λ : Kubelka-Munk optical concentration function at wavelength λ
K：光吸收系数K: light absorption coefficient
S：光散射系数S: light scattering coefficient
Rλ：波长λ时的反射率Rλ : reflectivity at wavelength λ
λ：波长λ: wavelength
另外，Duncan混色理论式如同以下所示。In addition, the Duncan color mixing theory is as shown below.
数4Number 4
Km：混色后的光吸收系数Km : light absorption coefficient after color mixing
Sm：混色后的光散射系数Sm : Light scattering coefficient after color mixing
Ki：着色剂i的光吸收系数Ki : light absorption coefficient of colorant i
Si：着色剂i的光散射系数Si : light scattering coefficient of colorant i
Pi：着色剂i的配合比率Pi : The mixing ratio of colorant i
Kubelka-Munk光学浓度式从分光反射率计算并求出光吸收系数与光散射系数的比，为了使用Duncan混色理论式进行混色计算，需要分别先求出光吸收系数及光散射系数。作为求出光吸收系数及光散射系数的方法，可使用公知的方法，例如可使用相对法或绝对法。The Kubelka-Munk optical concentration formula calculates the ratio of the light absorption coefficient to the light scattering coefficient from the spectral reflectance. In order to use the Duncan color mixing theory formula for color mixing calculation, it is necessary to first calculate the light absorption coefficient and the light scattering coefficient. As a method for calculating the light absorption coefficient and the light scattering coefficient, a known method can be used, such as a relative method or an absolute method.
此时，为了进一步提高预测精度，为了修正在形成涂料的树脂层与空气层的界面上发生的内部镜面反射或折射率差对分光反射率的测定产生的影响，能够在使用桑德森式变换成理想状态的反射率之后进行混色计算。另外，为了使着色剂的配合与目标色以致，在调整着色剂的配合比的方法中，可使用基于牛顿迭代法的反复计算，在对目标反射率与预测反射率的色彩一致性的评价中，可利用从反射率计算的色彩值XYZ、L＊a＊b＊等，可使用在对目标值与预测值差进行评价的同时利用牛顿迭代法进行收敛计算的metameric法，或者在目标反射率与预测反射率差的平方和进行评价的同时进行收敛计算的isomeric法。At this time, in order to further improve the prediction accuracy, in order to correct the influence of the internal mirror reflection or refractive index difference occurring at the interface between the resin layer and the air layer forming the coating on the measurement of the spectral reflectance, the color mixing calculation can be performed after the reflectance is converted into an ideal state using the Sanderson formula. In addition, in order to make the colorant matching the target color, in the method of adjusting the colorant matching ratio, iterative calculation based on the Newton iteration method can be used, and in the evaluation of the color consistency between the target reflectance and the predicted reflectance, the color value XYZ, L*a*b* calculated from the reflectance can be used, and the metameric method that uses the Newton iteration method to perform convergence calculation while evaluating the difference between the target value and the predicted value, or the isomeric method that performs convergence calculation while evaluating the sum of the squares of the difference between the target reflectance and the predicted reflectance can be used.
模糊推论可使用，用模糊集合论中的隶属函数定义含糊性的方法。作为模糊推论的具体方法，目前为止提出了各种方法，本发明中可使用任意方法。例如，可使用由曼达尼(Mamdani)研究出的模糊推论方法。Fuzzy inference can use a method of defining ambiguity using a membership function in fuzzy set theory. As a specific method of fuzzy inference, various methods have been proposed so far, and any method can be used in the present invention. For example, a fuzzy inference method developed by Mamdani can be used.
当S106工序中取得预测色彩数据Xa1时，可切换使用人工智能模型或者人工智能模型以外的预测式或者并用两者。When the predicted color data Xa1 is acquired in step S106, it is possible to switch between using an artificial intelligence model or a prediction formula other than an artificial intelligence model, or to use both.
本发明中，能够将S106工序做成只使用至少1种人工智能模型而取得预测色彩数据Xa1的工序。另外，在第2次以后的S106工序中，可做成并不使用人工智能模型而取得预测色彩数据Xa1的工序。In the present invention, step S106 can be a step of acquiring predicted color data Xa1 using only at least one artificial intelligence model. Also, in the second and subsequent steps of step S106, predicted color data Xa1 can be acquired without using an artificial intelligence model.
作为S106工序中可得到的预测色彩数据Xa1，可例举记录在所述数据库中的多种多样色彩数据。本发明中，优选预测色彩数据Xa1包含多角度的分光反射率及/或光亮感参数。通过使预测色彩数据Xa1包含多角度的分光反射率及/或光亮感参数，即使对于较难预测光学特性的光亮性色彩，也能够更加高精度地进行调色。As the predicted color data Xa1 that can be obtained in step S106, various color data recorded in the database can be cited. In the present invention, it is preferred that the predicted color data Xa1 include spectral reflectance and/or glossiness parameters at multiple angles. By making the predicted color data Xa1 include spectral reflectance and/or glossiness parameters at multiple angles, even for glossiness colors whose optical characteristics are difficult to predict, color matching can be performed with higher accuracy.
作为在合格与否的判定中的合格基准，例如可例举预测色彩数据Xa1与所述色彩数据Xp同等的情况。作为判断成同等的基准，例如能够分别单独对构成色彩数据Xp的各要素与构成预测色彩数据Xa1的各要素进行对比，而判断其差是否处于规定的范围内。例如，当色彩数据Xp及预测色彩数据Xa1包含使用L＊a＊b＊表色系的要素及从其得到的要素时，在L＊、a＊及b＊各自的基础上，还可以对色差ΔE也进行对比而判定合格与否。此时，还可以对各构成要素中的差分、一致度、误差率等设置阈值或者使用各种修正式，参考这些而由作业者、计算机或仪器，优选由计算机或仪器判定合格与否。As a pass criterion in the pass/fail judgment, for example, the case where the predicted color data Xa1 is equal to the color data Xp can be cited. As a criterion for judging the equivalence, for example, each element constituting the color data Xp can be compared with each element constituting the predicted color data Xa1 separately to judge whether the difference is within a specified range. For example, when the color data Xp and the predicted color data Xa1 include elements using the L*a*b* color system and elements obtained therefrom, the color difference ΔE can also be compared on the basis of L*, a* and b* to judge pass/fail. At this time, thresholds can be set for the difference, consistency, error rate, etc. in each constituent element, or various correction formulas can be used, and the operator, computer or instrument, preferably the computer or instrument, can refer to these to judge pass/fail.
本发明中，当所述预测色彩数据Xa1为光亮性所涉及的色彩数据时，优选使用多角度的分光反射率及/或光亮感参数判定合格与否。In the present invention, when the predicted color data Xa1 is color data related to glossiness, it is preferred to use multi-angle spectral reflectance and/or glossiness parameters to determine whether the data is acceptable or not.
并且，当判定合格与否时，根据需要还可以无论合格与否如何均可以通知作业者用于使预测色彩数据Xa1接近目标色彩数据Xp的改善点等。Furthermore, when judging whether the result is acceptable or not, the operator may be informed of improvement points for bringing the predicted color data Xa1 closer to the target color data Xp, if necessary, regardless of whether the result is acceptable or not.
S107工序S107 process
S107工序是当在所述S106工序中不合格时，使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，将从色彩数据Xp预测的不同于到此为止的预测配合组成数据的预测配合组成数据Yai，作为将所述组合物C1～Cn的1种以上作为成分的配合组成数据而取得，之后使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，取得从预测配合组成数据Yai预测的预测色彩数据Xai，同时达到合格为止反复进行通过与所述色彩数据Xp的比较而判定合格与否的工序。The S107 process is a process in which, when the S106 process fails, predicted combination composition data Yai predicted from the color data Xp and different from the predicted combination composition data so far is obtained as combination composition data having one or more of the compositions C1 to Cn as components using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, and then predicted color data Xai predicted from the predicted combination composition data Yai is obtained using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, and the process of repeatedly determining whether the data is qualified by comparing the data with the color data Xp is repeated until the data is qualified.
S107工序中，使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，将从色彩数据Xp预测的不同于到此为止的预测配合组成数据的预测配合组成数据Yai，作为将所述组合物C1～Cn的1种以上作为成分的配合组成数据而取得，作为这样的方法例如可例举以下的方法(1)～(5)。In the S107 process, using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, predicted combination composition data Yai predicted from the color data Xp, which is different from the predicted combination composition data so far, is obtained as combination composition data having one or more of the compositions C1 to Cn as components. As such a method, for example, the following methods (1) to (5) can be cited.
(1)通过使用多标签分类，将在满足色彩数据Xp的多个配合组成数据中未在S105中被选择的配合组成数据作为新的预测配合组成数据Yai，由此得到将所述组合物C1～Cn的1种以上作为成分的配合组成数据。(1) By using multi-label classification, the combination composition data not selected in S105 among the plurality of combination composition data satisfying the color data Xp is used as new predicted combination composition data Yai, thereby obtaining combination composition data having one or more of the compositions C1 to Cn as components.
(2)在所述组合物C1～Cn中改变可使用的种类数量等，通过导入在S105工序中未被使用的参数而从色彩数据Xp进行预测，将新的预测配合组成数据Yai作为所述组合物C1～Cn的1种以上为成分的配合组成数据而得到。(2) The number of types that can be used in the compositions C1 to Cn is changed, and a prediction is made from the color data Xp by importing parameters that are not used in the S105 process, and new predicted combination composition data Yai is obtained as combination composition data with one or more components of the compositions C1 to Cn.
(3)使用在S105工序中未被使用的人工智能模型，将从色彩数据Xp预测的不同于到此为止的预测配合组成数据的预测配合组成数据Yai，作为以所述组合物C1～Cn的1种以上为成分的配合组成数据而得到。(3) Using the artificial intelligence model not used in step S105, predicted combination composition data Yai predicted from the color data Xp and different from the predicted combination composition data thus far is obtained as combination composition data having one or more of the compositions C1 to Cn as components.
(4)使用在S105工序中未被使用的人工智能模型以外的预测式，将从色彩数据Xp预测的不同于到此为止的预测配合组成数据的预测配合组成数据Yai，作为以所述组合物C1～Cn的1种以上为成分的配合组成数据而得到。(4) Using a prediction formula other than the artificial intelligence model not used in step S105, predicted combination composition data Yai predicted from the color data Xp, which is different from the predicted combination composition data so far, is obtained as combination composition data having one or more of the compositions C1 to Cn as components.
(5)考虑在S106工序中进行的Xp与Xa1的比较得到的差分，使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，由此将预测配合组成数据Yai作为以所述组合物C1～Cn的1种以上为成分的配合组成数据而得到。(5) Taking into account the difference obtained by comparing Xp and Xa1 performed in step S106, the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model is used to obtain the predicted combination composition data Yai as combination composition data having one or more of the compositions C1 to Cn as components.
并且，使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，将从色彩数据Xp预测的预测配合组成数据Ya1，作为以所述组合物C1～Cn的1种以上为成分的配合组成数据而得到的方法，可与S105工序中的方法实质上相同。Furthermore, a method of obtaining the predicted combination composition data Ya1 predicted from the color data Xp as combination composition data having one or more of the compositions C1 to Cn as components using the learned artificial intelligence model and/or prediction formulas other than the artificial intelligence model can be substantially the same as the method in the S105 process.
另外，即使在S107工序中，也与S105工序同样地包含如下程序，即使用多标签分类将从色彩数据Xp预测的预测配合组成数据Ya1，作为以所述组合物C1～Cn的1种以上为成分的配合组成数据而得到。Also in step S107 , similarly to step S105 , a procedure is included for obtaining predicted compound composition data Ya1 predicted from color data Xp as compound composition data having one or more of the compositions C1 to Cn as components using multi-label classification.
另外，S107工序中，使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，取得从预测配合组成数据Yai预测的预测色彩数据Xai，同时通过与所述色彩数据Xp的比较判定合格与否，该方法相同于S106工序中的使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，取得从预测配合组成数据Ya1预测的预测色彩数据Xa1，同时通过与所述色彩数据Xp的比较判定合格与否的方法。In addition, in process S107, the predicted color data Xai predicted from the predicted combination composition data Yai is obtained by using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, and the pass or fail is determined by comparing it with the color data Xp. This method is the same as the method in process S106, in which the predicted color data Xa1 predicted from the predicted combination composition data Ya1 is obtained by using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, and the pass or fail is determined by comparing it with the color data Xp.
S107工序还可是反复进行微修正的工序，以便预测色彩数据Xa1与所述色彩数据Xp同等。The step S107 may be a step of repeatedly performing micro corrections so that the predicted color data Xa1 is equal to the color data Xp.
S107工序中，还可以设置有切换工序(手段)，以便只使用所述已学习的人工智能模型或人工智能模型以外的预测式的任意一方。本发明中，优选使用与当不合格时被使用的预测式不同的预测式。例如，当S106工序中使用特定的已学习的人工智能模型而得到的预测色彩数据Xai为不合格时，在接下来的S107工序中，优选切换使用特定的人工智能模型以外的预测式。In the step S107, a switching step (means) may be provided so as to use only one of the learned artificial intelligence model or the prediction formula other than the artificial intelligence model. In the present invention, it is preferred to use a prediction formula different from the prediction formula used when the result is unqualified. For example, when the predicted color data Xai obtained by using a specific learned artificial intelligence model in the step S106 is unqualified, in the next step S107, it is preferred to switch to using a prediction formula other than the specific artificial intelligence model.
可手动切换预测式，另外还可以设定为，当满足规定的条件时自动切换。本发明中，优选多次S107工序中的至少1次中使用人工智能模型以外的预测式。The prediction formula can be switched manually, or can be set to switch automatically when a predetermined condition is satisfied. In the present invention, it is preferred that a prediction formula other than the artificial intelligence model is used in at least one of the multiple S107 steps.
能够将在所述S107工序中得到的预测配合组成数据Yai作为所述组合物C1～Cn的1种以上的组合物的数据。另外，还可以作为树脂、色料、溶剂等各构成要素的配合量比的数据。优选考虑实际配合时的作业性等，而作为所述组合物C1～Cn的1种以上的组合物的数据。The predicted compounding composition data Yai obtained in the step S107 can be used as data of one or more compositions of the compositions C1 to Cn. In addition, it can also be used as data of the compounding amount ratio of each component such as resin, colorant, solvent, etc. It is preferred to use it as data of one or more compositions of the compositions C1 to Cn in consideration of the workability during actual compounding.
能够将在所述S107工序中得到的预测配合组成数据Yai作为所述组合物C1～Cn的1种以上的组合物的数据。另外，还可以作为树脂、色料、溶剂等各构成要素的配合量比的数据。优选考虑实际配合时的作业性等，而作为所述组合物C1～Cn的1种以上的组合物的数据。The predicted compounding composition data Yai obtained in the step S107 can be used as data of one or more compositions of the compositions C1 to Cn. In addition, it can also be used as data of the compounding amount ratio of each component such as resin, colorant, solvent, etc. It is preferred to use it as data of one or more compositions of the compositions C1 to Cn in consideration of the workability during actual compounding.
在S107工序中得到的预测配合组成数据Yai中，虽然并不特意限制所述组合物C1～Cn的种类，但是考虑实际配合时的作业性等，而可做成15种以下，优选做成12种以下，更优选做成10种以下。此时，作为含有金属颜料的组合物，做成5种以下，优选做成3种以下，作为含有珠光颜料的组合物，做成5种以下，优选做成3种以下。In the predicted compounding composition data Yai obtained in step S107, the types of the compositions C1 to Cn are not particularly limited, but in consideration of workability during actual compounding, the types may be 15 or less, preferably 12 or less, and more preferably 10 or less. In this case, the composition containing the metallic pigment may be 5 or less, preferably 3 or less, and the composition containing the pearlescent pigment may be 5 or less, preferably 3 or less.
S108工序S108 process
S108工序是当在所述S106或S107工序的任意一个中合格时，取得合格配合组成数据Yap1的工序。本发明中，既可以输出合格配合组成数据Yap1，还可以并不输出而发送数据。Step S108 is a step of acquiring the qualified combination composition data Yap1 when the result is qualified in either step S106 or step S107. In the present invention, the qualified combination composition data Yap1 may be output or the data may be transmitted without being output.
合格配合组成数据Yap1可以包含通过调色方法得到的涂料组合物的配合组成数据。例如，可例举多个市场上销售的调色用涂料的配合比以及颜料等的色料成分与调色用涂料的配合比以及1种以上的色料的配合比等的数据。The qualified composition data Yap1 may include the composition data of the coating composition obtained by the coloring method, for example, the composition ratio of a plurality of coloring coatings sold on the market, the composition ratio of colorant components such as pigments and the coloring coating, and the composition ratio of one or more colorants.
另外，合格配合组成数据Yap1可以包含用于消除合格配合组成与预测配合组成的差分所需的成分及/或关于其配合量的数据。例如，可例举在对合格配合组成与预测配合组成进行对比时的1以上的配合成分的差分等的1种以上的数据。这些差分数据相当于在从特定的配合组成进行微调色时使用的微调色配合组成数据，有助于简化调色作业。In addition, the qualified matching composition data Yap1 may include the components and/or data on the matching amount required to eliminate the difference between the qualified matching composition and the predicted matching composition. For example, one or more data such as the difference of more than one matching component when comparing the qualified matching composition with the predicted matching composition can be cited. These differential data are equivalent to the fine-tuning color matching composition data used when fine-tuning the color from a specific matching composition, which helps to simplify the color matching operation.
当输出合格配合组成数据Yap1时，可使用监视器、显示器、便携式终端装置、智能手机等便携式电话及根据信号可显示或输出信息或图像的任意的输出装置。另外，还可以使用根据信号能够将信息或图像显示于纸、塑料等适当的介质的印刷装置等输出装置。When outputting the qualified matching composition data Yap1, a monitor, display, portable terminal device, a portable phone such as a smart phone, and any output device that can display or output information or images according to a signal can be used. In addition, an output device such as a printing device that can display information or images on a suitable medium such as paper or plastic according to a signal can also be used.
合格配合组成数据Yap1的输出还可以是计算机内部的输出，此时，在计算机内部输出的合格配合组成数据Yap1通过通信手段等发送至自动配合装置、终端装置、数据记录装置、数据记录介质等。The output of the qualified matching composition data Yap1 can also be the output inside the computer. At this time, the qualified matching composition data Yap1 output inside the computer is sent to the automatic matching device, terminal device, data recording device, data recording medium, etc. through communication means.
另外，还可以并不输出合格配合组成数据Yap1，而是通过通信手段等发送至自动配合装置、终端装置、数据记录装置、数据记录介质等。In addition, instead of outputting the qualified combination composition data Yap1, it is also possible to send it to an automatic combination device, a terminal device, a data recording device, a data recording medium, etc. through a communication means.
S109工序S109 process
S109工序是根据所述合格配合组成数据Yap1调制实际候补涂料CMap1，得到该实际候补涂料CMap1的涂装板而取得实测色彩数据Xap1的工序。Step S109 is a step of modulating the actual candidate paint CMap1 based on the acceptable blend composition data Yap1, obtaining a painted plate of the actual candidate paint CMap1, and acquiring the measured color data Xap1.
并不特意限定实际候补涂料CMap1的调制方法，而是可通过调制涂料时的公知的方法进行。例如，可以将构成实际候补涂料CMap1的各成分放入调和容器，根据需要通过搅拌装置或分散装置等进行混合而进行调制。另外，可以作为原色涂料而混合1种以上市场上销售的组合物而进行调制。The preparation method of the actual candidate paint CMap1 is not particularly limited, and it can be prepared by a known method for preparing paint. For example, the components constituting the actual candidate paint CMap1 can be put into a mixing container and mixed by a stirring device or a dispersing device as needed to prepare it. In addition, it can be prepared by mixing one or more commercially available compositions as primary color paints.
本发明中，还可以将由计算机算出的所述合格配合组成数据Yap1经由有线或无线网络发送至具备电子天平等的自动调和机，由此调制实际候补涂料CMap1。由此，即使作业者并不是熟练者，也能够容易地调制实际候补涂料CMap1。In the present invention, the qualified combination composition data Yap1 calculated by the computer can be sent to an automatic mixer equipped with an electronic balance via a wired or wireless network to prepare the actual candidate paint CMap1. Thus, even if the operator is not a skilled worker, the actual candidate paint CMap1 can be easily prepared.
关于取得实际候补涂料CMap1的涂装板的方法，也并不特意进行限定，而是可通过调制涂装板时的公知的方法进行。例如，可例举如下地形成涂装板的方法等，基材上以隐蔽膜厚以上的方式形成1层以上的调色涂料的涂膜，最上层上例如以10～100μm干燥膜厚度的膜厚形成透明涂料的涂膜。当形成各涂膜时，还可以根据需要进行加热而实施干燥、固化。当通过加热实施干燥、固化时，既可以在全部涂膜已形成之后统一进行，还可以在每当形成涂膜时进行。实际候补涂料CMap1的涂装板既可以使用采用机器人等的自动涂装装置全自动制作，还可以由作业者进行一部分工序而制作。The method for obtaining the coating plate of the actual candidate paint CMap1 is not particularly limited, and can be carried out by a known method for modulating the coating plate. For example, the following method of forming a coating plate can be cited, wherein a coating film of more than one layer of a tinting paint is formed on a substrate in a manner greater than a concealing film thickness, and a coating film of a clear paint is formed on the top layer with a film thickness of, for example, 10 to 100 μm dry film thickness. When each coating film is formed, it is also possible to heat and dry and cure as needed. When drying and curing are performed by heating, it can be performed uniformly after all coating films have been formed, or it can be performed each time a coating film is formed. The coating plate of the actual candidate paint CMap1 can be produced fully automatically using an automatic coating device using a robot or the like, or it can be produced by an operator performing a portion of the process.
作为在本发明方法中使用的基材，并不特意进行限定，而是可使用用于制作调色用试验涂板的基材。例如，可例举金属板、纸、塑料薄膜等。只要基材的大小为可进行测色且可目视确认色调程度的大小。则并不特意限制，例如通常为1个边的长度为5～20cm左右。The substrate used in the method of the present invention is not particularly limited, and a substrate used for preparing a test coating plate for color matching can be used. For example, metal plates, paper, plastic films, etc. can be exemplified. As long as the size of the substrate is such that color measurement can be performed and the degree of color tone can be visually confirmed, it is not particularly limited, for example, the length of one side is usually about 5 to 20 cm.
对实际候补涂料CMap1的涂装板进行测色而取得实测色彩数据Xap1的工序，通过使用色彩计、多角度分光光度计、激光式金属感测定仪器、变角分光光度计、光泽计、微观光亮感测定仪等测定仪器的测定直接取得，或者可使用通过测定取得的数据进行算出而取得。The process of measuring the color of the coated plate of the actual candidate paint CMap1 to obtain the actual measured color data Xap1 can be directly obtained by measuring using a measuring instrument such as a colorimeter, a multi-angle spectrophotometer, a laser metallic sense measuring instrument, a variable angle spectrophotometer, a gloss meter, a microscopic gloss sense measuring instrument, etc., or can be obtained by calculation using the data obtained through the measurement.
S110工序S110 process
S110工序是通过所述色彩数据Xp与所述实测色彩数据Xap1的比较及/或作为所述目标的色彩与所述实际候补涂料CMap1的涂装板的色彩的比较，判定合格与否的工序。通过作业者、计算机或仪器判定合格与否。Step S110 is a step of determining whether the color data Xp is acceptable by comparing the color data Xp with the measured color data Xap1 and/or comparing the target color with the color of the plate coated with the actual candidate paint CMap1. The determination of whether the color is acceptable is made by an operator, a computer or an instrument.
由作业者通过目视对具有目标色彩数据Xp的物品(涂板)的色彩等与所述实际候补涂料CMap1的涂装板的色彩进行对比，由此能够判定合格与否。另外，例如还可以分别单独对目标色彩数据Xp或构成其的各要素与所述实际候补涂料CMap1的涂装板的通过色度计等进行测定而得到的色彩数据或构成其的各要素进行对比，而由作业者、计算机或仪器判定合格与否。此时，与S106工序及S107工序中的合格与否的判定同样，还可以对各构成要素中的差分、一致度、误差率等设置阈值或者使用各种修正式，参考这些而由作业者、计算机或仪器判定合格与否。The operator visually compares the color of the object (coated plate) having the target color data Xp with the color of the coated plate of the actual candidate paint CMap1, thereby determining whether it is acceptable or not. In addition, for example, the target color data Xp or each element constituting it can be separately compared with the color data or each element constituting it obtained by measuring the coated plate of the actual candidate paint CMap1 by a colorimeter, etc., and the operator, computer or instrument determines whether it is acceptable or not. At this time, similar to the determination of the acceptance or not in the S106 process and the S107 process, thresholds can be set for the difference, consistency, error rate, etc. in each component, or various correction forms can be used, and the operator, computer or instrument can refer to these to determine whether it is acceptable or not.
本发明中，当计算机判定合格与否时，例如可采用机器学习。例如，可使用选自梯度提升的决策树、线形回归、逻辑回归、简单感知器、MLP、神经网络、支持向量机、随机森林、高斯过程、贝叶斯网络、k近邻法、其他机器学习中使用的模型的1种以上。In the present invention, when the computer determines whether the test is qualified or not, for example, machine learning can be used. For example, one or more models selected from gradient boosted decision trees, linear regression, logistic regression, simple perceptron, MLP, neural network, support vector machine, random forest, Gaussian process, Bayesian network, k nearest neighbor method, and other machine learning models can be used.
本发明中，优选采用选自使用神经网络、梯度提升的决策树及高斯过程的1种以上，更优选采用选自使用神经网络及梯度提升的决策树的1种以上。In the present invention, it is preferred to use one or more selected from the group consisting of a neural network, a decision tree using gradient boosting, and a Gaussian process, and it is more preferred to use one or more selected from the group consisting of a neural network and a decision tree using gradient boosting.
本发明中，可采用神经网络中的SOM(Self-Organizing Map：自组织映射网络图)。In the present invention, a SOM (Self-Organizing Map) in a neural network may be used.
此时，通过SOM追加独自的数据解释(例如，SOM网络图上设定矢量，而定义判定方向)而进行改善，由此对学习中使用的数据可提高判定精度。In this case, the SOM is improved by adding its own interpretation of the data (for example, setting vectors on the SOM network diagram to define the determination direction), thereby improving the determination accuracy of the data used in the learning.
另外，优选对SOM其本身的算法追加改善，在SOM的并不存在数据的节点，预先定义从周围节点推断出的矢量，由此也能够应对未知数据。由此，通过SOM将色及金属感类似的彼此节点(色彩)配置在SOM网络图上的附近，由此关于具有学习数据及未知数据的色彩(颜色及金属感)，能够高精度地判定一致度的合格与否。In addition, it is preferred to improve the algorithm of the SOM itself, and predefine vectors inferred from surrounding nodes at nodes of the SOM where no data exists, so that unknown data can also be handled. Thus, the SOM arranges nodes (colors) with similar colors and metallic senses near each other on the SOM network diagram, so that the pass or fail of the consistency of colors (colors and metallic senses) with learning data and unknown data can be determined with high accuracy.
本发明中，为了减轻作业者的负担，优选通过计算机或仪器判定合格与否。In the present invention, in order to reduce the burden on the operator, it is preferred that the pass/fail determination be made by a computer or an instrument.
本发明中，还可以组合基于作业者的目视的合格与否的判定与基于色彩数据或构成其的各要素由作业者、计算机或仪器进行的合格与否的判定来进行合格与否的判定。In the present invention, the pass/fail judgment can be performed by combining the pass/fail judgment based on the operator's visual inspection with the pass/fail judgment based on the color data or each element constituting the color data by the operator, computer or instrument.
并且，当判定合格与否时，根据需要还可以无论合格与否如何均可以通知作业者关于实际候补涂料CMap1的配合组成的改善点等。Furthermore, when judging whether the paint is qualified or not, the operator may be informed of points for improvement of the blending composition of the actual candidate paint CMap1 as needed, regardless of whether the paint is qualified or not.
当S110工序中合格时，可根据实际候补涂料CMap1的配合组成调制涂料。另外，根据需要，可以不使用计算机，而是附加由作业者进行微调色的工序，进一步接近作为目标的色彩。If the process S110 is qualified, the paint can be prepared according to the actual composition of the candidate paint CMap 1. In addition, as required, a process of fine-tuning the color by the operator can be added without using a computer to further approach the target color.
S111工序S111 process
S111工序是当在所述S110工序中不合格时，达到合格为止将所述S105～S110工序或S107～S110工序反复进行的工序。The step S111 is a step in which, when the product fails in the step S110, the steps S105 to S110 or the steps S107 to S110 are repeated until the product passes.
本发明中，即使当S110工序中的合格与否的判定中2次以上不合格时，可反复进行所述S105～S110工序或S107～S110工序，S110工序中的合格与否的判定中合格为止反复进行。In the present invention, even if the pass/fail determination in step S110 fails twice or more, the steps S105 to S110 or steps S107 to S110 may be repeated until the pass/fail determination in step S110 passes.
当反复进行所述S105～S110工序或S107～S110工序时，S105工序～S107工序中，与前一次同样，可使用所述已学习的人工智能模型及/或人工智能模型以外的预测式而取得预测色彩数据Xa1或Xai。另外，还可以设置有切换工序(手段)，以便只使用所述已学习的人工智能模型或人工智能模型以外的预测式的任意一方。本发明中，当使用至少1种已学习的人工智能模型而得到的预测色彩数据Xa1或Xai为不合格时，优选切换成下一次使用人工智能模型以外的预测式。When the steps S105 to S110 or the steps S107 to S110 are repeatedly performed, in the steps S105 to S107, the predicted color data Xa1 or Xai can be obtained by using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, as in the previous time. In addition, a switching step (means) can be provided so as to use only one of the learned artificial intelligence model or the prediction formula other than the artificial intelligence model. In the present invention, when the predicted color data Xa1 or Xai obtained by using at least one learned artificial intelligence model is unqualified, it is preferred to switch to a prediction formula other than the artificial intelligence model for the next use.
当S111工序中并不合格而反复进行所述S105～S110工序或S107～S110工序时，当S105工序～S107工序中取得预测色彩数据Xa1或Xai时，可切换成在所述已学习的人工智能模型及/或人工智能模型以外的预测式中使用上一次并未使用的已学习的人工智能模型及/或预测式。可手动切换，另外还可以设定为，当满足规定的条件时自动切换。本发明中，优选多次S105工序～S107工序中的至少1次中使用人工智能模型以外的预测式。When the S111 process fails and the S105-S110 process or the S107-S110 process is repeated, when the predicted color data Xa1 or Xai is obtained in the S105-S107 process, the learned artificial intelligence model and/or prediction formula that was not used last time can be switched to the predicted artificial intelligence model and/or prediction formula other than the learned artificial intelligence model. Manual switching can be performed, and it can also be set to automatically switch when the specified conditions are met. In the present invention, it is preferred that a prediction formula other than the artificial intelligence model is used in at least one of the multiple S105-S107 processes.
另外，当S111工序中不合格时，优选将预测色彩数据Xa1或Xai与实测色彩数据的差分Δ作为修正系数α而输入到计算机，之后反复进行所述S105～S110工序或S107～S110工序。In addition, when the S111 process fails, it is preferred to input the difference Δ between the predicted color data Xa1 or Xai and the measured color data into the computer as a correction coefficient α, and then repeat the S105 to S110 process or the S107 to S110 process.
由此，在实际候补涂料的调制次数的5次以内，优选3次以内，更优选2次以内，能够调制出可得到作为目的的色彩的涂料。Thus, a paint capable of obtaining a target color can be prepared within 5 times, preferably within 3 times, and more preferably within 2 times of the number of preparations of the actual candidate paint.
基于计算机调色的涂料的制作方法(本发明的第2实施形态)Method for producing paint based on computer color matching (second embodiment of the present invention)
图7是当执行基于本发明的第2实施形态所涉及的计算机调色的涂料的制作方法时的流程图。并且，图7所示的流程只不过是本发明的一个实施方式。Fig. 7 is a flow chart of a method for producing a paint by computer coloring according to a second embodiment of the present invention. The flow chart shown in Fig. 7 is merely one embodiment of the present invention.
基于本发明的第2实施形态所涉及的计算机调色的涂料的制作方法是使用具备数据库和计算机的装置且包括下述S201～S211工序的方法，在该数据库中至少登记有1种以上的组合物的色彩数据X及配合组成数据Y，在该计算机中利用登记在该数据库中的数据的配色计算逻辑发挥作用。The method for preparing a computer-colored paint according to the second embodiment of the present invention is a method using a device having a database and a computer and including the following steps S201 to S211, wherein at least one color data X and a combination composition data Y of a composition are registered in the database, and a color matching calculation logic based on the data registered in the database is used in the computer.
以下，对S201～S211工序进行详细说明。Hereinafter, steps S201 to S211 will be described in detail.
S201工序S201 process
S201工序是使用登记在所述数据库中的数据，向所述计算机输入学习用数据的工序。Step S201 is a step of inputting learning data into the computer using the data registered in the database.
S201工序中使用的学习用数据，可以与本发明的第1实施形态所涉及的在基于计算机调色的涂料的制作方法中的S101工序中所使用的学习数据相同。The learning data used in step S201 may be the same as the learning data used in step S101 in the method for producing a paint by computer color matching according to the first embodiment of the present invention.
S201工序中的向计算机的输入手段，可以与本发明的第1实施形态所涉及的在基于计算机调色的涂料的制作方法中的S101工序中的向计算机的输入手段相同。The input means to the computer in step S201 may be the same as the input means to the computer in step S101 in the method for producing the paint by computer color matching according to the first embodiment of the present invention.
S202工序S202 process
S202工序是使所述学习用数据进行机器学习，生成包含从配合组成数据Y推断色彩数据X的人工智能模型的至少1种的已学习的人工智能模型的工序。作为本发明中的人工智能模型，例如可由选自使用梯度提升的决策树、线形回归、逻辑回归、简单感知器、MLP、神经网络、支持向量机、随机森林、高斯过程、贝叶斯网络、k近邻法、其他机器学习中使用的模型的1种以上所构成。本发明中，优选由选自使用神经网络、梯度提升的决策树及高斯过程的1种以上所构成，尤其优选采用选自使用神经网络及梯度提升的决策树的1种以上的人工智能模型。Step S202 is a step of subjecting the learning data to machine learning to generate at least one learned artificial intelligence model including an artificial intelligence model for inferring color data X from matching composition data Y. The artificial intelligence model in the present invention may be, for example, a decision tree using gradient boosting, linear regression, logistic regression, simple perceptron, MLP, neural network, support vector machine, random forest, Gaussian process, Bayesian network, k-nearest neighbor method, or one or more models used in other machine learning. In the present invention, it is preferably composed of one or more selected from decision trees using neural networks, gradient boosting, and Gaussian processes, and it is particularly preferred to use one or more artificial intelligence models selected from decision trees using neural networks and gradient boosting.
本发明中，由选自使用神经网络、梯度提升的决策树及高斯过程的1种以上所构成，使用S201工序中被输入的学习用数据而使神经网络学习，由此能够生成包含从配合组成数据Y推断色彩数据X的已学习的人工智能模型的至少1种的已学习的人工智能模型。In the present invention, it is composed of one or more selected from the group consisting of a neural network, a decision tree using gradient boosting, and a Gaussian process, and the neural network is learned using the learning data input in the S201 process, thereby being able to generate at least one learned artificial intelligence model that includes a learned artificial intelligence model for inferring color data X from the combination composition data Y.
使用S201工序中输入到计算机的学习用数据，实现人工智能模型(神经网络)的学习。学习用数据至少使用1种以上的组合物所涉及的色彩数据X和组成配合数据Y。作为神经网络的算法，可使用有教师学习方法之一即公知的误差逆传播算法。设定表示学习速度的参数即学习率(0～1之间的实数值)及学习中的输出值的误差的容许值即容许误差(0～1之间的实数值)，而使神经网络学习。由此，能够将配合组成数据Y所涉及的1种以上的特征量与基于该配合组成数据Y的涂料的涂膜的色彩数据X所涉及的1种以上的特征量关联起来。使用已被学习的网络，通过前馈计算能够预测满足色彩数据的组成配合数据及基于组成配合数据的色彩数据。已被学习的网络，并不进行费用或时间等工时所涉及的实验性确认而预测这些。The learning data input into the computer in the S201 process is used to realize the learning of the artificial intelligence model (neural network). The learning data uses at least one color data X and composition data Y related to the composition. As the algorithm of the neural network, one of the teacher learning methods, namely the well-known error back propagation algorithm, can be used. The parameters representing the learning speed, namely the learning rate (real value between 0 and 1) and the allowable value of the error of the output value during learning, namely the allowable error (real value between 0 and 1), are set to make the neural network learn. In this way, it is possible to associate one or more feature quantities related to the combination composition data Y with one or more feature quantities related to the color data X of the coating film of the paint based on the combination composition data Y. Using the learned network, the composition data that satisfies the color data and the color data based on the composition data can be predicted by feedforward calculation. The learned network predicts these without experimental confirmation involving cost or time and other working hours.
本发明中，生成所述已学习的人工智能模型的至少1种的工序，可以与本发明的第1实施形态所涉及的在基于计算机调色的涂料的制作方法中的S102工序的生成已学习的人工智能模型的至少1种的工序相同。In the present invention, the process of generating at least one of the learned artificial intelligence models may be the same as the process of generating at least one of the learned artificial intelligence models in the S102 process in the method for producing a paint based on computer color matching involved in the first embodiment of the present invention.
S203工序S203 process
S203工序是取得作为目标的色彩的目标色彩数据Xt的工序。Step S203 is a step of acquiring target color dataXt of a target color.
作为目标色彩数据Xt，可例举关于涂装物、成形品、自然结构物等所具有的全部色彩的色彩数据。尤其，优选作为涂装物的色彩数据。The target color data Xt may be color data on all colors of painted objects, molded products, natural structures, etc. In particular, color data on painted objects is preferred.
本发明即使将目前为止难以进行计算机调色的含有光亮性颜料的涂膜的色彩数据作为目标色彩数据Xt，也能够高精度地进行调色。因此，优选S203工序中的目标色彩数据Xt为含有光亮性颜料的涂膜的色彩数据。当然，S203工序中的目标色彩数据Xt还可以为并不含有光亮性颜料的涂膜的色彩数据。The present invention can perform color matching with high accuracy even if the color data of a coating film containing a bright pigment, which has been difficult to color match by computer, is used as the target color data Xt . Therefore, it is preferred that the target color data Xt in step S203 is the color data of a coating film containing a bright pigment. Of course, the target color data Xt in step S203 may also be the color data of a coating film that does not contain a bright pigment.
构成目标色彩数据Xt的要素，可以与构成登记在数据库中的色彩数据的要素相同。例如，可以为通过计量仪器测定的色彩数据或从其算出的色彩数据。The elements constituting the target color dataXt may be the same as the elements constituting the color data registered in the database, for example, color data measured by a measuring instrument or color data calculated therefrom.
作为用于取得色彩数据的计量仪器，只要是可测定光亮涂膜(金属涂膜、珍珠色涂膜等)、纯色涂膜等的色彩而取得色彩数据的计量仪器，则并不特意限制采用测定原理、测定值的色彩数据的算出方法等，而是可使用现有公知的计量仪器。例如，可使用具备照射被测色表面的光源的单角度分光光度计、多角度分光光度计、色彩计、色差计、变角分光光度计等色度计及摄像装置、微观光亮感测定仪等测定仪器及色样本卡等计量仪器的1个以上。另外，可任意使用对从这些计量仪器得到的各种色彩数据进行处理的数据处理装置。As a measuring instrument for obtaining color data, as long as it can measure the color of a bright coating (metallic coating, pearl coating, etc.), a solid color coating, etc. and obtain color data, there is no particular limitation on the measurement principle adopted, the method of calculating the color data of the measured value, etc., and any known measuring instrument can be used. For example, a colorimeter such as a single-angle spectrophotometer, a multi-angle spectrophotometer, a colorimeter, a colorimeter, a variable-angle spectrophotometer, and a camera device, a measuring instrument such as a microscopic brightness meter, and a color sample card can be used. In addition, a data processing device that processes various color data obtained from these measuring instruments can be used arbitrarily.
作业者使用各种计量仪器，直接测定被测色物来可得到目标色彩数据Xt。另外，还可以由各种计量仪器根据程序等自动取得。而且，还可以根据这些测色数据进行算出。The operator can obtain the target color data Xt by directly measuring the color object using various measuring instruments. Alternatively, the target color data X t can be automatically obtained by various measuring instruments according to a program, etc. Furthermore, the target color data X t can be calculated based on the color measurement data.
本发明中，优选使用多角度分光光度计对被测色表面进行测定，而取得目标色彩数据Xt。In the present invention, it is preferred to use a multi-angle spectrophotometer to measure the color surface to be measured, and obtain the target color data Xt .
另外，当目标色彩数据Xt并不是直接测定被测色物而得到的数据时，能够将从被测色物的商品名等得到的色彩数据作为目标色彩数据Xt而加以使用。例如，当目标色彩数据Xt为有关汽车的色彩数据时，可根据汽车的商品名、型号、年式、制造号等得到的涂料数据设定目标色彩数据Xt。In addition, when the target color dataXt is not data obtained by directly measuring the color object to be measured, color data obtained from the product name of the color object to be measured, etc. can be used as the target color dataXt . For example, when the target color dataXt is color data related to a car, the target color dataXt can be set based on paint data obtained from the product name, model, year, manufacturing number, etc. of the car.
S204工序S204 process
S204工序是向所述计算机输入所述目标色彩数据Xt的工序。Step S204 is a step of inputting the target color dataXt into the computer.
S204工序中的向计算机的输入手段，可以与本发明的第1实施形态所涉及的在基于计算机调色的涂料的制作方法中的S104工序中的向计算机的输入手段相同。The input means to the computer in step S204 may be the same as the input means to the computer in step S104 in the method for producing the paint by computer color matching according to the first embodiment of the present invention.
S205工序S205 process
S205工序是通过使用计算机的检索，取得近似于所述目标色彩数据Xt的检索色彩数据Xn1及对应于检索色彩数据Xn1的近似配合组成数据Yn1，同时对所述目标色彩数据Xt与所述检索色彩数据Xn1进行比较而判定合格与否的工序。Step S205 is a step of obtaining search color dataXn1 similar to the target color dataXt and approximate combination composition dataYn1 corresponding to the search color dataXn1 by searching with a computer, and comparing the target color dataXt with the search color dataXn1 to determine whether they are acceptable.
S205工序可以作为相当于计算机选色(CCS)的工序，能够在登记于数据库的多个色彩数据中，检索近似于目标色彩数据Xt的色彩数据，作为检索色彩数据Xn1而取得。The step S205 may be a step corresponding to computer color selection (CCS), and can search for color data similar to the target color dataXt among a plurality of color data registered in the database and obtain the searched color dataXn1 .
在此，登记于数据库的色彩数据例如为公知的色样本账的色彩数据或过去制作的涂板的色彩数据等，均与对应于色彩数据的配合组成数据有关联。从而，通过取得检索色彩数据Xn1，将对应的配合组成数据即近似配合组成数据Yn1也能够容易得到。Here, the color data registered in the database, such as the color data of a known color sample book or the color data of a previously produced paint plate, are all associated with the matching composition data corresponding to the color data. Therefore, by obtaining the search color dataXn1 , the corresponding matching composition data, i.e., the approximate matching composition dataYn1, can also be easily obtained.
分别对构成色彩数据的要素的1个以上(例如，L＊a＊b＊表色系中的各值等)，与构成目标色彩数据Xt的对应的要素进行对比，能够通过检索值的差分、一致度、误差率等处于一定的范围内的要素来取得检索色彩数据Xn1。所述一定的范围既可以由作业者参考经验等来设定，另外还可以通过计算机来设定。By comparing one or more elements constituting the color data (e.g., each value in the L*a*b* colorimetric system) with the corresponding elements constituting the target color dataXt , the search color dataXn1 can be obtained by searching for elements whose difference, consistency, error rate, etc. of the search value are within a certain range. The certain range can be set by the operator based on experience or by a computer.
S205工序中，当比较目标色彩数据Xt与所述检索色彩数据Xn1而判定合格与否时，可着眼于构成检索色彩数据Xn1的要素的1个以上与构成目标色彩数据Xt的要素的1个以上，对分别对应的各构成要素进行比较来进行。当判定合格与否时，例如还可以对各构成要素中的差分、一致度、误差率等设置阈值，参考这些由仪器或作业者判定合格与否。此时，还可以反映熟练作业者的观点等，而在各构成要素之间进行加权。In step S205, when comparing the target color dataXt with the search color dataXn1 to determine whether the data is acceptable or not, the comparison can be performed by focusing on one or more elements constituting the search color dataXn1 and one or more elements constituting the target color dataXt , and comparing the corresponding constituent elements. When determining whether the data is acceptable or not, for example, a threshold value can be set for the difference, consistency, error rate, etc. in each constituent element, and the device or the operator can determine whether the data is acceptable or not by referring to these. At this time, the opinions of the skilled operator can be reflected, and weighting can be performed between the constituent elements.
S206工序S206 process
S206工序是当在所述S205工序中不合格时，使用计算机取得预测为提供目标色彩数据Xt的候补配合组成数据Yni，之后使用所述至少1种已学习的人工智能模型及/或所述人工智能模型以外的预测式，取得从候补配合组成数据Yni预测的预测色彩数据Xni，同时对所述色彩数据Xt与所述预测色彩数据Xni进行比较而判定合格与否的工序。Step S206 is a step of obtaining candidate combination composition data Yni predicted to provide target color data Xt using a computer when the step S205 fails, then obtaining predicted color data Xni predicted from the candidate combination composition data Yni using the at least one learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, and comparing the color data Xt with the predicted color data Xni to determine whether the color data X t is qualified or not.
作为使用计算机取得预测为提供目标色彩数据Xt的候补配合组成数据Yni的方法，例如可例举作为计算机色彩校正(CCM)而已周知的方法，即基于使用计算机的配色计算逻辑的算出或基于数理最优化的算出。As a method of obtaining candidate combination composition dataYni predicted to provide target color dataXt using a computer, for example, a method known as computer color correction (CCM), that is, calculation based on color matching calculation logic using a computer or calculation based on mathematical optimization can be cited.
基于使用计算机的配色计算逻辑的算出，例如基于登记在所述数据库中的各种色彩数据以及与其对应的组成配合数据，对目标色彩数据Xt与所述各种色彩数据进行比较而以差分、一致度等处于一定的范围的方式进行计算，由此将认为最合理的一个以上的配合组成决定为候补配合组成数据Yni。利用构成计算逻辑的各种函数，通过较小的反复工序可修正任意的配合组成或近似配合组成。此时，以有规则的形态生成理论指令，能够补助计算速度或调整算法的精度。Based on the calculation of the color matching calculation logic using a computer, for example, based on the various color data registered in the database and the corresponding composition matching data, the target color dataXt is compared with the various color data and the calculation is performed in a manner that the difference, consistency, etc. are within a certain range, thereby determining one or more matching compositions that are considered to be the most reasonable as candidate matching composition dataYni . Using various functions that constitute the calculation logic, any matching composition or approximate matching composition can be corrected through a small iterative process. At this time, generating theoretical instructions in a regular form can supplement the calculation speed or adjust the accuracy of the algorithm.
候补配合组成数据Yni例如参照记录在数据库中的色彩数据，对每一个构成通过近似配合组成得到的色彩数据的坐标轴，检索具有在减小误差的方向上发挥作用的特性信息的成分，由此可得到通过数理最优化的算出。例如，L＊a＊b＊表色系中，当将误差作为ΔL＊＝L＊2-L＊1、Δa＊＝a＊2-a＊1、Δb＊＝b＊2-b＊1时，当L＊轴上的误差ΔL＊为正值时，检索具有在减小L＊2值的方向上发挥作用的特性信息的成分，当L＊轴上的误差ΔL＊为负值时，检索具有增加L＊2值的特性信息的成分。同样地，分别当a＊轴上误差Δa＊为正值时，检索具有减小a＊2值的特性信息(绿)的成分，当a＊轴上的误差Δa＊为负值时，检索具有增加a＊2值的特性信息(红)的成分，当b＊轴上的误差Δb＊为正值时，检索具有减小b＊2值的特性信息(青)的成分，当b＊轴上的误差Δb＊为负值时，检索具有增加b＊2值的特性信息(黄)的成分。由此，近似配合组成中，对每一个构成色空间的表色系的坐标轴，通过追加在减小误差的方向上发挥作用即用于赋予规定的特性信息的成分，能够取得接近作为目标的色彩的候补配合组成数据Yni。The candidate combination composition dataYni is calculated by mathematical optimization by, for example, referring to the color data recorded in the database, searching for a component having characteristic information that acts in a direction of reducing the error for each coordinate axis constituting the color data obtained by the approximate combination composition. For example, in the L*a*b* color system, when the error is ΔL*=L*2 -L*1 , Δa*=a*2 -a*1 , Δb*=b*2 -b*1 , when the error ΔL* on the L* axis is a positive value, searching for a component having characteristic information that acts in a direction of reducing the L*2 value, and when the error ΔL* on the L* axis is a negative value, searching for a component having characteristic information that increases the L*2 value. Similarly, when the error Δa* on the a* axis is a positive value, a component having characteristic information (green) that reduces the a*2 value is retrieved, when the error Δa* on the a* axis is a negative value, a component having characteristic information (red) that increases the a*2 value is retrieved, when the error Δb* on the b* axis is a positive value, a component having characteristic information (cyan) that reduces the b*2 value is retrieved, and when the error Δb* on the b* axis is a negative value, a component having characteristic information (yellow) that increases the b*2 value is retrieved. Thus, in the approximate combination composition, by adding a component that acts in the direction of reducing the error, i.e., for providing a predetermined characteristic information, to each coordinate axis of the color system constituting the color space, candidate combination composition data Yni close to the target color can be obtained.
假设，当并未检索到具有在减小误差的方向上发挥作用的特性信息的成分时，根据目标色彩数据Xt，能够在得到更适合的新的近似配合组成之后取得候补配合组成数据。Assuming that a component having characteristic information that works in a direction of reducing the error is not retrieved, candidate combination composition data can be acquired after a more suitable new approximate combination composition is obtained based on the target color data Xt .
另外，还可以对通过CCM得到的配合组成数据进行作业者的修正(例如，公知的色样本账的色彩数据、过去制作的涂板的色彩数据、参考自身经验等而进行的作业者的修正)、所述计算机的修正、人工智能模型的修正等而得到候补配合组成数据Yni。In addition, the combination composition data obtained through CCM can also be corrected by the operator (for example, color data of a known color sample account, color data of a coating plate made in the past, corrections made by the operator based on his own experience, etc.), corrections by the computer, corrections by the artificial intelligence model, etc. to obtain the candidate combination composition data Yni .
而且，在汽车修理工场等的作业现场中，当可使用的色料或组合物被限制时等，还可以只根据作业现场中可使用的色料或组合物而取得候补配合组成数据Yni。Furthermore, when the colorants or compositions that can be used are limited at a work site such as an automobile repair shop, the candidate blend composition data Yni can be acquired based on only the colorants or compositions that can be used at the work site.
可通过显示手段或印刷手段等输出候补配合组成数据Yni。另外，还可以并不输出，而是从计算机发送至实施下一个工序的仪器等。The candidate combination composition data Yni may be outputted by display means or printing means, etc. Alternatively, the candidate combination composition data Y ni may be transmitted from the computer to a device or the like that performs the next step instead of being outputted.
当S206工序中取得预测色彩数据Xni时，可使用人工智能模型及/或人工智能模型以外的预测式。可切换使用人工智能模型或者人工智能模型以外的预测式。When the predicted color dataXni is obtained in step S206, an artificial intelligence model and/or a prediction formula other than the artificial intelligence model may be used. The artificial intelligence model or a prediction formula other than the artificial intelligence model may be switched for use.
本发明中，能够将S206工序做成只使用至少1种人工智能模型而取得预测色彩数据Xni的工序。另外，在第2次以后的S206工序中，可做成并不使用人工智能模型而取得预测色彩数据Xni的工序。In the present invention, the step S206 can be a step of obtaining the predicted color dataXni using only at least one artificial intelligence model. In addition, in the second and subsequent steps S206, the predicted color dataXni can be obtained without using an artificial intelligence model.
作为S206工序中可得到的预测色彩数据Xni，可例举记录在所述数据库中的多种多样色彩数据。本发明中，优选预测色彩数据Xni包含多角度的分光反射率及/或光亮感参数。通过使预测色彩数据Xni包含多角度的分光反射率及/或光亮感参数，即使对于较难预测光学特性的光亮性色彩，也能够更加高精度地进行调色。As the predicted color dataXni that can be obtained in step S206, various color data recorded in the database can be cited. In the present invention, it is preferred that the predicted color dataXni include spectral reflectance and/or glossiness parameters at multiple angles. By making the predicted color dataXni include spectral reflectance and/or glossiness parameters at multiple angles, even for glossiness colors whose optical characteristics are difficult to predict, color matching can be performed with higher accuracy.
S206工序中，作为使用已被学习的人工智能模型取得预测色彩数据Xni的方法，向已被学习的人工智能模型的神经网络中的输入层的各单元，输入候补配合组成数据Yni的特征量即可。输入到输入层的候补配合组成数据Yni，在各节点及各层之间一边被加权一边被发送，从输出层的各单元作为色彩数据而被输出。In step S206, as a method of obtaining predicted color dataXni using the learned artificial intelligence model, the feature amount of candidate combination composition dataYni is input to each unit of the input layer in the neural network of the learned artificial intelligence model. The candidate combination composition dataYni input to the input layer is transmitted while being weighted between each node and each layer, and is output as color data from each unit of the output layer.
S206工序中，作为使用已被学习的人工智能模型以外的预测式取得预测色彩数据Xni的方法，可使用利用CCM的调色领域中的公知的各种预测式。作为这样的预测式例如可例举采用基于Kubelka-Munk光学浓度式和Duncan混色理论式的2定数法的预测式的方法，采用模糊推论的方法，以及其他通过计算机对色彩数据或配合组成数据进行比较而将各自的匹配程度指数化的方法等。In step S206, as a method of obtaining the predicted color dataXni using a prediction formula other than the learned artificial intelligence model, various prediction formulas known in the field of color matching using CCM can be used. Examples of such prediction formulas include a method using a prediction formula based on the Kubelka-Munk optical density formula and the Duncan color mixing theory formula, a method using fuzzy inference, and other methods of comparing color data or matching composition data by computer and indexing the degree of matching.
采用Kubelka-Munk光学浓度式和Duncan混色理论式的方法，可以与本发明的第1实施形态所涉及的在基于计算机调色的涂料的制作方法中的S106工序中所记载的使用Kubelka-Munk光学浓度式和Duncan混色理论式的方法相同。The method using the Kubelka-Munk optical concentration formula and the Duncan color mixing theory formula can be the same as the method using the Kubelka-Munk optical concentration formula and the Duncan color mixing theory formula described in the S106 process in the computer-based color matching paint production method involved in the first embodiment of the present invention.
采用模糊推论的方法，可以与本发明的第1实施形态所涉及的在基于计算机调色的涂料的制作方法中的S106工序中所记载的使用模糊推论的方法相同。The method using fuzzy inference may be the same as the method using fuzzy inference described in step S106 in the method for producing a paint by computer color matching according to the first embodiment of the present invention.
例如，通过分别单独对构成色彩数据Xt的各要素与构成预测色彩数据Xni的各要素进行对比，由此能够判定合格与否。例如，当色彩数据Xt及预测色彩数据Xni包含使用L＊a＊b＊表色系的要素及从其得到的要素时，在L＊、a＊及b＊各自的基础上，还可以对色差ΔE也进行对比而判定合格与否。此时，还可以对各构成要素中的差分、一致度、误差率等设置阈值，参考其而判定合格与否。For example, by comparing each element constituting the color dataXt with each element constituting the predicted color dataXni , it is possible to determine whether the data is acceptable or not. For example, when the color dataXt and the predicted color dataXni include elements using the L*a*b* color system and elements obtained therefrom, the color difference ΔE may be compared on the basis of L*, a*, and b* to determine whether the data is acceptable or not. In this case, a threshold value may be set for the difference, consistency, error rate, etc. in each component element, and the threshold value may be used to determine whether the data is acceptable or not.
本发明中，当所述预测色彩数据Xni为光亮性所涉及的色彩数据时，优选使用多角度的分光反射率及/或光亮感参数判定合格与否。In the present invention, when the predicted color dataXni is color data related to glossiness, it is preferred to use multi-angle spectral reflectance and/or glossiness parameters to determine whether it is acceptable or not.
并且，当判定合格与否时，根据需要还可以无论合格与否如何均可以通知作业者用于使预测色彩数据Xni接近目标色彩数据Xt的改善点等。Furthermore, when judging whether the result is acceptable or not, the operator may be informed of improvement points for bringing the predicted color data Xni closer to the target color data Xt , regardless of whether the result is acceptable or not, as necessary.
S207工序S207 process
S207工序是当在所述S206工序中不合格时，使用计算机取得预测为提供目标色彩数据Xt的候补配合组成数据Yni，之后使用所述至少1种已学习的人工智能模型及/或所述人工智能模型以外的预测式，取得从候补配合组成数据Yni预测的预测色彩数据Xni，同时对所述色彩数据Xt与所述预测色彩数据Xni进行比较，达到合格为止将判定合格与否的工序反复进行的工序。Step S207 is a step of obtaining candidate combination composition data Yni predicted to provide target color data Xt using a computer when the step S206 fails, then obtaining predicted color data X ni predicted from the candidate combination composition data Yni using the at least one learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, and comparing the color data Xt with the predicted color data Xni , and repeating the process of judging whether the color data X t is qualified until the color data Xt is qualified.
S207工序中，使用计算机取得预测为提供目标色彩数据Xt的候补配合组成数据Yni的方法，可以与S206工序中的使用计算机取得预测为提供目标色彩数据Xt的候补配合组成数据Yni的方法相同。In step S207, the method of using a computer to obtain candidate combination composition data Yni predicted to provide target color data Xt may be the same as the method of using a computer to obtain candidate combination composition data Yni predicted to provide target color data Xt in step S206.
另外，S207工序中，判定合格与否的方法与S206工序中的判定合格与否的方法相同。In addition, the method of judging whether the product is good or bad in step S207 is the same as the method of judging whether the product is good or bad in step S206.
S207工序中，当取得从候补配合组成数据Yni预测的预测色彩数据Xni时，可使用与S206工序相同的至少1种已学习的人工智能模型及/或人工智能模型以外的预测式。另外，还可以设置有切换工序(手段)，以便只使用至少1种已学习的人工智能模型或人工智能模型以外的预测式的任意一方。本发明中，优选使用与当不合格时被使用的预测式不同的预测式。例如，当S206工序中使用至少1种已学习的人工智能模型而得到的预测色彩数据Xni为不合格时，在接下来的S207工序中，优选切换使用人工智能模型以外的预测式。In the step S207, when the predicted color dataXni predicted from the candidate combination composition dataYni is obtained, at least one of the learned artificial intelligence models and/or the prediction formula other than the artificial intelligence model can be used as in the step S206. In addition, a switching step (means) can be provided so that only one of the at least one learned artificial intelligence model or the prediction formula other than the artificial intelligence model is used. In the present invention, it is preferred to use a prediction formula different from the prediction formula used when the result is unqualified. For example, when the predicted color dataXni obtained by using at least one learned artificial intelligence model in the step S206 is unqualified, it is preferred to switch to use the prediction formula other than the artificial intelligence model in the next step S207.
可手动切换预测式，另外还可以设定为，当满足规定的条件时自动切换。本发明中，优选多次S207工序中的至少1次中使用人工智能模型以外的预测式。The prediction formula can be switched manually, or can be set to switch automatically when a predetermined condition is satisfied. In the present invention, it is preferred that a prediction formula other than the artificial intelligence model is used in at least one of the multiple S207 steps.
S208工序S208 process
S208工序是当在所述S205～S207工序的任意一个中合格时，取得合格配合组成数据YC1的工序。本发明中，既可以输出合格配合组成数据YC1，还可以并不输出而发送数据。Step S208 is a step of acquiring the qualified combination composition data YC1 when any of the steps S205 to S207 is qualified. In the present invention, the qualified combination composition data YC1 may be output or the data may be transmitted without being output.
合格配合组成数据YC1可以包含通过调色方法得到的涂料组合物的配合组成数据。例如，可例举多个市场上销售的调色用涂料的配合比以及颜料等的色料成分与调色用涂料的配合比以及1种以上的色料的配合比等的数据。The qualified composition data YC1 may include the composition data of the coating composition obtained by the coloring method, for example, the composition ratio of a plurality of coloring coatings sold on the market, the composition ratio of coloring components such as pigments and the coloring coating, and the composition ratio of one or more colorants.
另外，合格配合组成数据YC1可以包含用于消除合格配合组成与近似配合组成及/或候补配合组成的差分所需的成分及/或关于其配合量的数据。例如，可例举在对合格配合组成与近似配合组成或候补配合组成进行对比时的1以上的配合成分的差分以及在对近似配合组成与候补配合组成进行对比时的1以上的配合成分的差分等的1种以上的数据。这些差分数据相当于在从特定的配合组成进行微调色时使用的微调色配合组成数据，有助于简化调色作业。In addition, the qualified combination composition data YC1 may include data on the components and/or the amounts thereof required to eliminate the difference between the qualified combination composition and the approximate combination composition and/or the candidate combination composition. For example, the data may include the difference of more than one combination component when comparing the qualified combination composition with the approximate combination composition or the candidate combination composition, and the difference of more than one combination component when comparing the approximate combination composition with the candidate combination composition. These difference data are equivalent to the fine-tuning color combination composition data used when fine-tuning the color from a specific combination composition, which helps to simplify the color adjustment operation.
当输出合格配合组成数据YC1时，可使用监视器、显示器、便携式终端装置、智能手机等便携式电话及根据信号可显示或输出信息或图像的任意的输出装置。另外，还可以使用根据信号能够将信息或图像显示于纸、塑料等适当的介质的印刷装置等输出装置。When outputting the qualified matching composition data YC1 , a monitor, display, portable terminal device, a mobile phone such as a smart phone, and any output device that can display or output information or images according to a signal can be used. In addition, an output device such as a printing device that can display information or images on a suitable medium such as paper or plastic according to a signal can also be used.
合格配合组成数据YC1的输出还可以是计算机内部的输出，此时，在计算机内部输出的合格配合组成数据YC1通过通信手段等发送至自动配合装置、终端装置、数据记录装置、数据记录介质等。The output of the qualified combination composition data YC1 can also be the output inside the computer. In this case, the qualified combination composition data Y C1output inside the computer is sent to the automatic combination device, terminal device, data recording device, data recording medium, etc. through communication means.
另外，还可以并不输出合格配合组成数据YC1，而是通过通信手段等发送至自动配合装置、终端装置、数据记录装置、数据记录介质等。Furthermore, instead of outputting the qualified combination composition data YC1 , it is also possible to transmit it to an automatic combination device, a terminal device, a data recording device, a data recording medium, or the like through a communication means or the like.
S209工序S209 process
S209工序是根据所述合格配合组成数据YC1调制实际候补涂料CMCi，得到该实际候补涂料CMCi的涂装板而取得实测色彩数据XCi的工序。Step S209 is a step of preparing the actual candidate paint CMCi based on the acceptable blend composition data YC1 , obtaining a painted plate of the actual candidate paint CMCi , and acquiring the measured color data XCi .
并不特意限定实际候补涂料CMCi的调制方法，而是可通过调制涂料时的公知的方法进行。例如，可以将构成实际候补涂料CMCi的各成分放入调和容器，根据需要通过搅拌装置或分散装置等进行混合而进行调制。The preparation method of the actual candidate paint CMCi is not particularly limited, and it can be prepared by a known method for preparing paint. For example, the components constituting the actual candidate paint CMCi can be put into a mixing container and mixed by a stirring device or a dispersing device as needed.
本发明中，还可以将由计算机算出的合格配合组成数据YC1经由有线或无线网络发送至具备电子天平等的自动调和机，由此调制实际候补涂料CMCi。由此，即使作业者并不是熟练者，也能够容易地调制实际候补涂料CMCi。In the present invention, the qualified compound composition dataYC1 calculated by the computer can be sent to an automatic mixer equipped with an electronic balance via a wired or wireless network to prepare the actual candidate paint CMCi . Thus, even if the operator is not a skilled worker, the actual candidate paint CMCi can be easily prepared.
关于取得实际候补涂料CMCi的涂装板的方法，也并不特意进行限定，而是可通过调制涂装板时的公知的方法进行。例如，可例举如下地形成涂装板的方法等，基材上以隐蔽膜厚以上的方式形成1层以上的调色涂料的涂膜，最上层上例如以10～100μm干燥膜厚度的膜厚形成透明涂料的涂膜。当形成各涂膜时，还可以根据需要进行加热而实施干燥、固化。当通过加热实施干燥、固化时，既可以在全部涂膜已形成之后统一进行，还可以在每当形成涂膜时进行。实际候补涂料CMCi的涂装板既可以使用采用机器人等的自动涂装装置全自动制作，还可以由作业者进行一部分工序而制作。The method for obtaining the coating plate of the actual candidate paint CMCi is not particularly limited, but can be carried out by a known method when modulating the coating plate. For example, the following method of forming a coating plate can be cited, wherein a coating film of more than one layer of toning paint is formed on the substrate in a manner greater than the concealing film thickness, and a coating film of a clear paint is formed on the top layer with a film thickness of, for example, 10 to 100 μm dry film thickness. When each coating film is formed, it is also possible to heat and dry and cure as needed. When drying and curing are performed by heating, it can be performed uniformly after all the coating films have been formed, or it can be performed each time the coating film is formed. The coating plate of the actual candidate paint CMCi can be made fully automatically using an automatic coating device using a robot, etc., or it can be made by an operator performing a part of the process.
作为在本发明方法中使用的基材，并不特意进行限定，而是可使用用于制作调色用试验涂板的基材。例如，可例举金属板、纸、塑料薄膜等。只要基材的大小为可进行测色且可目视确认色调程度的大小。则并不特意限制，例如通常为1个边的长度为5～20cm左右。The substrate used in the method of the present invention is not particularly limited, and a substrate used for preparing a test coating plate for color matching can be used. For example, metal plates, paper, plastic films, etc. can be exemplified. As long as the size of the substrate is such that color measurement can be performed and the degree of color tone can be visually confirmed, it is not particularly limited, for example, the length of one side is usually about 5 to 20 cm.
对实际候补涂料CMCi的涂装板进行测色而取得实测色彩数据XCi的工序，通过使用色彩计、多角度分光光度计、激光式金属感测定仪器、变角分光光度计、光泽计、微观光亮感测定仪等测定仪器的测定直接取得，或者可使用通过测定取得的数据进行算出而取得。The step of measuring the color of the coated plate of the actual candidate paint CMCi to obtain the actual color data XCi is directly obtained by measurement using a measuring instrument such as a colorimeter, a multi-angle spectrophotometer, a laser metallographic measuring instrument, a variable angle spectrophotometer, a gloss meter, a microscopic gloss meter, or the like, or is obtained by calculation using data obtained by the measurement.
S210工序S210 process
S210工序是通过所述色彩数据Xt与所述实测色彩数据XCi的比较及/或作为所述目标的色彩与所述实际候补涂料CMCi的涂装板的色彩的比较，判定合格与否的工序。通过作业者、计算机或仪器判定合格与否。Step S210 is a step of determining whether the color dataXt is acceptable by comparing the color data Xt with the measured color dataXci and/or comparing the target color with the color of the plate coated with the actual candidate paintCMci . The determination of whether the color is acceptable is performed by an operator, a computer or an instrument.
由作业者通过目视对作为目标的色彩Xt与所述实际候补涂料CMCi的涂装板的色彩进行对比，由此能够判定合格与否。另外，例如还可以分别单独对目标色彩数据Xt或构成其的各要素与所述实际候补涂料CMCi的涂装板的通过色度计等进行测定而得到的色彩数据或构成其的各要素进行对比，而由作业者、计算机或仪器判定合格与否。此时，与S206工序及S207工序中的合格与否的判定同样，还可以对各构成要素中的差分、一致度、误差率等设置阈值或者使用各种修正式，参考这些而由作业者、计算机或仪器判定合格与否。The operator visually compares the target colorXt with the color of the coating plate of the actual candidate paint CMCi , thereby determining whether it is acceptable or not. In addition, for example, the target color dataXt or each element constituting it can be separately compared with the color data or each element constituting it obtained by measuring the coating plate of the actual candidate paint CMCi by a colorimeter, etc., and the operator, computer or instrument determines whether it is acceptable or not. At this time, similar to the determination of the acceptance or not in the S206 process and the S207 process, a threshold value can be set for the difference, consistency, error rate, etc. in each component, or various correction forms can be used, and the operator, computer or instrument can refer to these to determine whether it is acceptable or not.
本发明中，当由计算机判定合格与否时，可以与本发明的第1实施形态所涉及的在基于计算机调色的涂料的制作方法中的S110工序中所记载的合格与否的判定相同。In the present invention, when the pass/fail determination is made by a computer, the pass/fail determination may be the same as the pass/fail determination described in step S110 in the method for producing a paint by computer color matching according to the first embodiment of the present invention.
本发明中，为了减轻作业者的负担，优选通过计算机或仪器判定合格与否。In the present invention, in order to reduce the burden on the operator, it is preferred that the pass/fail determination be made by a computer or an instrument.
本发明中，还可以组合基于作业者的目视的合格与否的判定与基于色彩数据或构成其的各要素由作业者、计算机或仪器进行的合格与否的判定来进行合格与否的判定。In the present invention, the pass/fail judgment can be performed by combining the pass/fail judgment based on the operator's visual inspection with the pass/fail judgment based on the color data or each element constituting the color data by the operator, computer or instrument.
并且，当判定合格与否时，根据需要还可以无论合格与否如何均可以通知作业者关于实际候补涂料CMCi的配合组成的改善点等。Furthermore, when judging whether the candidate is qualified or not, the operator can be informed of improvement points of the blending composition of the actual candidate paint CMCi , regardless of whether the candidate is qualified or not, as necessary.
当S210工序中合格时，可根据实际候补涂料CMCi的配合组成调制涂料。另外，根据需要，可以不使用计算机，而是附加由作业者进行微调色的工序，进一步接近作为目标的色彩。If the S210 process is qualified, the paint can be prepared according to the actual composition of the candidate paint CMCi . In addition, as needed, a process of fine-tuning the color by the operator can be added without using a computer to further approach the target color.
S211工序S211 process
S211工序是当在所述S210工序中的合格与否的判定中不合格时，反复进行所述S206～S210工序的工序。Step S211 is a step of repeating the steps S206 to S210 when the product fails the pass/fail determination in step S210.
本发明中，即使当S210工序中的合格与否的判定中2次以上不合格时，可反复进行所述S206～S210工序，S210工序中的合格与否的判定中合格为止反复进行。In the present invention, even if the pass/fail determination in step S210 fails two or more times, steps S206 to S210 may be repeatedly performed until the pass/fail determination in step S210 passes.
当反复进行S206～S210工序时，S206工序及/或S207工序中，与前一次同样，可使用至少1种已学习的人工智能模型及/或人工智能模型以外的预测式而取得预测色彩数据Xni。另外，还可以设置有切换工序(手段)，以便只使用至少1种已学习的人工智能模型或人工智能模型以外的预测式的任意一方。本发明中，当使用至少1种已学习的人工智能模型而得到的预测色彩数据Xni为不合格时，优选切换成下一次使用人工智能模型以外的预测式。When the steps S206 to S210 are repeatedly performed, in the step S206 and/or the step S207, as in the previous step, the predicted color data Xni can be obtained by using at least one of the learned artificial intelligence models and/or the prediction formula other than the artificial intelligence model. In addition, a switching step (means) can be provided so that only one of the at least one learned artificial intelligence model or the prediction formula other than the artificial intelligence model is used. In the present invention, when the predicted color data Xni obtained by using at least one learned artificial intelligence model is unqualified, it is preferred to switch to the prediction formula other than the artificial intelligence model for the next time.
当S211工序中并不合格而反复进行S206～S210工序时，当S206工序及/或S207工序中取得预测色彩数据Xni时，可切换成在至少1种已学习的人工智能模型及/或人工智能模型以外的预测式中使用上一次并未使用的预测式。可手动切换预测式，另外还可以设定为，当满足规定的条件时自动切换。本发明中，优选多次S206工序及/或S207工序中的至少1次中使用人工智能模型以外的预测式。When the process S211 fails and the processes S206 to S210 are repeated, when the predicted color dataXni is obtained in the process S206 and/or the process S207, a prediction formula that was not used last time can be used in at least one of the learned artificial intelligence models and/or prediction formulas other than the artificial intelligence model. The prediction formula can be switched manually, and can also be set to automatically switch when a specified condition is met. In the present invention, it is preferred that a prediction formula other than the artificial intelligence model is used in at least one of the multiple processes S206 and/or the process S207.
另外，当S211工序中不合格时，优选将预测色彩数据Xni与实测色彩数据XCi的差分Δ作为修正系数α而输入到计算机，之后反复进行S206工序～S211工序。When the result of step S211 fails, it is preferred that the difference Δ between the predicted color dataXni and the measured color dataXci is input into the computer as the correction coefficient α, and then the steps S206 to S211 are repeated.
由此，在S209工序中的实际候补涂料CMCi的调制次数的5次以内，优选3次以内，更优选2次以内，能够调制出可得到作为目的的色彩的涂料。Thus, a paint that can obtain a target color can be prepared within 5 times, preferably within 3 times, and more preferably within 2 times of the number of preparations of the actual candidate paint CMCi in step S209.
基于计算机调色的涂料的制作方法的应用Application of the method for making paint based on computer color matching
基于本发明的第1及第2实施形态所涉及的计算机调色的涂料的制作方法，均可在用于取得作为目的的色彩的涂料的调制中被使用。另外，可在涂料的配合组成的识别或配合组成的修正中被使用。The methods for producing computer-colored paint according to the first and second embodiments of the present invention can be used in the preparation of paint to obtain a desired color, or in the identification or correction of the composition of the paint.
尤其，基于本发明的第1及第2实施形态所涉及的计算机调色的涂料的制作方法，可在涂布于具有色彩的物品，例如汽车、摩托车等车辆或其零件及卡车、公交车、电车、单轨电车等大型车辆或其零件以及其他工业产品等的修补用涂料的调制中被使用。具有色彩的物品尤其还可以是具有单层或多层涂膜的物品。尤其在含有金属涂色、真珠光泽色等光亮性颜料的涂膜上设置有透明涂膜的多层涂膜时，可最大限地发挥本发明的效果。In particular, the method for producing computer-colored paint according to the first and second embodiments of the present invention can be used in the preparation of paint for repairing to be applied to colored articles, such as vehicles such as automobiles and motorcycles or their parts, large vehicles such as trucks, buses, trams, monorails or their parts, and other industrial products. The colored articles can also be articles having a single-layer or multi-layer coating film. In particular, when a multi-layer coating film having a transparent coating film is provided on a coating film containing bright pigments such as metallic coloring and pearlescent color, the effect of the present invention can be maximized.
预测涂膜的色彩数据的方法(本发明的第3实施形态)Method for predicting color data of coating film (third embodiment of the present invention)
图8是当执行本发明的第3实施形态所涉及的预测涂膜的色彩数据的方法时的流程图。并且，图8所示的流程只不过是本发明的一个实施方式。Fig. 8 is a flow chart of the method for predicting color data of a coating film according to the third embodiment of the present invention. The flow chart shown in Fig. 8 is merely one embodiment of the present invention.
本发明的第3实施形态所涉及的预测涂膜的色彩数据的方法是使用具备数据库和计算机的计算机调色装置且包括下述S301～S307工序的方法，在该数据库中至少登记有1种以上的组合物的色彩数据X及配合组成数据Y，在该计算机中利用登记在该数据库中的数据的配色计算逻辑发挥作用。The method for predicting the color data of a coating film according to the third embodiment of the present invention is a method using a computer color matching device having a database and a computer and including the following steps S301 to S307, wherein at least one color data X and a combination composition data Y of a composition are registered in the database, and a color matching calculation logic of the data registered in the database is used in the computer.
在此，S301工序及S302工序分别与S101工序及S102工序相同。以下，对S303～S307工序进行详细说明。Here, the steps S301 and S302 are the same as the steps S101 and S102, respectively. The steps S303 to S307 will be described in detail below.
S303工序S303 process
S303工序是取得预测涂膜的色彩数据的涂料CMt的配合组成数据YCM的工序。Step S303 is a step of acquiring the mixing composition data YCM of the paint CMt which is the color data of the predicted coating film.
将预测涂膜的色彩数据的涂料CMt，实际上并不调制涂料而能够作为想要取得其色彩或色彩数据的涂料。由此，例如在一次试制多个色彩等时，能够不进行涂料的调制、利用涂装的涂膜的调制及对涂膜的色彩数据的测定这样的工序，而能够容易取得色彩或色彩数据。The paintCMt of the color data of the coating film is predicted, and the color or color data of the coating film can be obtained without actually preparing the paint. Thus, for example, when multiple colors are tested at one time, the color or color data can be easily obtained without the steps of preparing the paint, preparing the coating film by painting, and measuring the color data of the coating film.
配合组成数据YCM是关于涂料中含有的粘合剂、着色颜料、添加剂成分等的各自种类(商品名、品号等)及配合量所涉及的数据。具体而言，可与构成登记在数据库中的配合组成数据的组成及配合量所涉及的数据相同。The compounding composition dataYCM is data on the types (trade names, product numbers, etc.) and compounding amounts of the binder, coloring pigment, additive components, etc. contained in the paint. Specifically, it can be the same as the data on the composition and compounding amount constituting the compounding composition data registered in the database.
S304工序S304 process
S304工序是向所述计算机输入所述配合组成数据YCM的工序。Step S304 is a step of inputting the combination composition data YCM into the computer.
数据的输入可以使用与所述S104工序中向计算机输入目标色彩数据Xt的手段相同的手段。The data can be inputted by the same means as the means for inputting the target color dataXt into the computer in the above-mentioned step S104.
S305工序S305 process
S305工序是根据需要，通过使用计算机的检索，取得对应于所述配合组成数据YCM的检索色彩数据Xn1的工序。The step S305 is a step of acquiring the search color dataXn1 corresponding to the mixed composition dataYCM by searching using a computer as needed.
S305工序可以是于计算机选色(CCS)相同的工序，从登记在数据库中的多个配合组成数据中，检索近似于所述配合组成数据YCM的配合组成数据，将对应于所得到的配合组成数据的色彩数据作为检索色彩数据Xn1而取得。Step S305 may be the same step as computer color selection (CCS), in which matching composition data similar to the matching composition data YCM are retrieved from a plurality of matching composition data registered in the database, and color data corresponding to the obtained matching composition data are obtained as retrieved color data Xn1 .
在此，登记于数据库的配合组成数据例如为市场上销售的涂料的配合组成数据或过去制作的涂料的配合组成数据等，均与对应于配合组成数据的色彩数据有关联。从而，通过取得近似于配合组成数据YCM的配合组成数据，将对应的色彩数据作为检索色彩数据Xn1而能够容易得到。Here, the combination composition data registered in the database, for example, the combination composition data of paint sold on the market or the combination composition data of paint produced in the past, are all associated with the color data corresponding to the combination composition data. Therefore, by obtaining the combination composition data similar to the combination composition data YCM , the corresponding color data can be easily obtained as the search color data Xn1 .
当从多个配合组成数据检索近似于配合组成数据YCM的配合组成数据时，分别对构成配合组成数据的要素的1个以上(例如，特定颜色的颜料的含量等)，与对应的要素进行对比，能够通过检索值的差分、一致度、误差率等处于一定的范围内的要素来取得。所述一定的范围既可以由作业者参考经验等来设定，另外还可以通过计算机来设定。When searching for combination composition data similar to combination composition data YCM from a plurality of combination composition data, one or more elements constituting the combination composition data (e.g., the content of a specific color pigment, etc.) are compared with the corresponding elements, and elements within a certain range of the difference, consistency, error rate, etc. of the search value can be obtained. The certain range can be set by the operator based on experience, etc., or can be set by a computer.
本发明中，根据需要执行S305工序，即使并不执行也不存在任何问题。In the present invention, the step S305 is executed as needed, but there is no problem even if it is not executed.
S306工序S306 process
S306工序是当在所述S305工序中未检索到对应的检索色彩数据Xn1时，或者当并未进行所述S305工序时，使用所述至少1种已学习的人工智能模型或所述至少1种已学习的人工智能模型和所述人工智能模型以外的预测式，从所述配合组成数据YCM取得预测色彩数据Xm1的工序。Step S306 is a step for obtaining predicted color dataXm1 from the combination composition data YCM using the at least one learned artificial intelligence model or the at least one learned artificial intelligence model and a prediction formula other than the artificial intelligence model when the corresponding search color dataXn1 is not retrieved in the stepS305 or when the step S305 is not performed.
本发明中，使用至少1种人工智能模型，能够从配合组成数据YCM取得预测色彩数据Xm1。另外，并用至少1种已学习的人工智能模型和人工智能模型以外的预测式，能够取得预测色彩数据Xm1。在此，使用人工智能模型的方法等、人工智能模型以外的预测式及使用其的方法等，可以与所述S106工序中已记载的内容相同。In the present invention, at least one artificial intelligence model is used to obtain the predicted color dataXm1 from the combination composition dataYCM . In addition, at least one learned artificial intelligence model and a prediction formula other than the artificial intelligence model are used to obtain the predicted color dataXm1 . Here, the method of using the artificial intelligence model, the prediction formula other than the artificial intelligence model and the method of using the same, etc. can be the same as those described in the above step S106.
S307工序S307 process
S307工序是根据需要，取得涂装有所述涂料CMt的涂装板的实测色彩数据XCM，与所述预测色彩数据Xm1进行比较的工序。Step S307 is a step of acquiring the actual color data XCM of the painted plate painted with the paint CMt as needed, and comparing the data with the predicted color data Xm1 .
通过实施S307工序且反馈结果，能够更高精度地预测涂膜的色彩数据。By performing the step S307 and feeding back the result, the color data of the coating film can be predicted with higher accuracy.
例如，当实测色彩数据XCM与预测色彩数据Xm1的背离较大时，通过只使用人工智能模型以外的预测式而重新取得预测色彩数据，同时与实测色彩数据XCM进行比较，由此能够进行反馈。For example, when the measured color data XCM and the predicted color data Xm1 differ greatly, feedback can be provided by reacquiring predicted color data using only a prediction formula other than the artificial intelligence model and comparing it with the measured color data XCM .
另外，还可以将预测色彩数据Xm1与实测色彩数据XCM的差分Δ作为修正系数β而输入到计算机，之后反复进行S305工序～S307工序。Alternatively, the difference Δ between the predicted color data Xm1 and the measured color data XCM may be input into a computer as a correction coefficient β, and then the steps S305 to S307 may be repeated.
预测涂膜的色彩数据的方法的应用Application of the method for predicting the color data of coating films
例如，当进行车辆涂装等涂装时，在调制涂料时，能够在涂膜的色调的预测中使用本发明的预测涂膜的色彩数据的方法。For example, when painting a vehicle or the like, the method of predicting color data of a coating film of the present invention can be used to predict the color tone of the coating film when preparing the coating material.
通过使用本发明预测涂膜的色彩数据的方法，不需要生成关于个别涂料分配组成的特定的涂料，可较高精度预测关于被指定的多个涂料分配组成的色彩。另外，在多个涂料分配组成候补中，能够容易选择从被指定的色彩偏离最小的涂料分配组成。由此，关于多个涂料分配组成不需要分别调制涂料，也不需要在此之后实际涂布于被涂布材而制作涂板之后进行测定，能够取得对应的色彩数据。By using the method of predicting the color data of the coating film of the present invention, it is not necessary to generate a specific coating for an individual coating distribution composition, and the colors of the specified multiple coating distribution compositions can be predicted with high accuracy. In addition, among multiple coating distribution composition candidates, the coating distribution composition with the smallest deviation from the specified color can be easily selected. Thus, it is not necessary to prepare the coatings for the multiple coating distribution compositions separately, nor is it necessary to actually apply them to the coated material and then make a coating plate for measurement, and the corresponding color data can be obtained.
计算机调色系统(本发明的第4实施形态)Computer color matching system (fourth embodiment of the present invention)
本发明的第4实施形态所涉及的计算机调色系统是具备数据库和计算机的计算机调色系统且包括下述手段S401～S411，在该数据库中登记有1种以上的组合物C1～Cn(n为2以上的整数)的各自配合组成数据Y1～Yn和分别对应于各配合组成数据的色彩数据X1～Xn，在该计算机中利用登记在该数据库中的数据的配色计算逻辑发挥作用。The computer color matching system involved in the fourth embodiment of the present invention is a computer color matching system having a database and a computer and includes the following means S401~S411, in which the respective matching composition data Y1~Yn of one or more compositions C1~Cn (n is an integer greater than 2) and the color data X1~Xn corresponding to each matching composition data are registered in the database, and the color matching calculation logic of the data registered in the database is functioned in the computer.
(S401)是使用登记在所述数据库中的数据，向所述计算机输入学习用数据的手段。(S401) is a means for inputting learning data into the computer using the data registered in the database.
(S402)是使所述学习用数据进行机器学习，生成包含从色彩数据X推断配合组成数据Y的人工智能模型的至少1种的已学习的人工智能模型的手段。(S402) is a means for subjecting the learning data to machine learning to generate at least one learned artificial intelligence model including an artificial intelligence model for inferring the combination composition data Y from the color data X.
(S403)是取得配合组成Yp为未知的作为目标的色彩的色彩数据Xp的手段。(S403) is a means for obtaining color data Xp of a target color whose matching composition Yp is unknown.
(S404)是向所述计算机输入所述色彩数据Xp的手段。(S404) is a means for inputting the color data Xp into the computer.
(S405)是使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，将从色彩数据Xp预测的预测配合组成数据Ya1，作为将所述组合物C1～Cn的1种以上作为成分的配合组成数据而取得的手段。(S405) is a means for obtaining the predicted combination composition data Ya1 predicted from the color data Xp as combination composition data having one or more of the compositions C1 to Cn as components, using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model.
(S406)是使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，取得从预测配合组成数据Ya1预测的预测色彩数据Xa1，同时与所述色彩数据Xp进行比较而判定合格与否的手段。(S406) is a means of using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model to obtain the predicted color data Xa1 predicted from the predicted combination composition data Ya1, and compare it with the color data Xp to determine whether it is qualified or not.
(S407)是当在所述手段S406中不合格时，使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，将从色彩数据Xp预测的不同于到此为止的预测配合组成数据的预测配合组成数据Yai，作为将所述组合物C1～Cn的1种以上作为成分的配合组成数据而取得，之后使用所述已学习的人工智能模型及/或人工智能模型以外的预测式，取得从预测配合组成数据Yai预测的预测色彩数据Xai，同时达到合格为止反复进行通过与所述色彩数据Xp的比较而判定合格与否的手段。(S407) is a method of, when the method S406 fails, using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, obtaining predicted combination composition data Yai predicted from the color data Xp, which is different from the predicted combination composition data so far, as combination composition data having one or more of the compositions C1 to Cn as components, and then using the learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, obtaining predicted color data Xai predicted from the predicted combination composition data Yai, and repeatedly determining whether the method is qualified by comparing the data with the color data Xp until the method is qualified.
(S408)是当在所述手段S406或S407的任意一个中合格时，取得合格配合组成数据Yap1的手段。(S408) is a means for obtaining the qualified combination composition data Yap1 when any one of the means S406 or S407 is qualified.
(S409)是根据所述合格配合组成数据Yap1调制实际候补涂料CMap1，得到该实际候补涂料CMap1的涂装板而取得实测色彩数据Xap1的手段。(S409) is a means for modulating the actual candidate paint CMap1 based on the qualified combination composition data Yap1, obtaining a painted plate of the actual candidate paint CMap1, and acquiring the measured color data Xap1.
(S410)是通过所述色彩数据Xp与所述实测色彩数据Xap1的比较及/或作为所述目标的色彩与所述实际候补涂料CMap1的涂装板的色彩的比较，判定合格与否的手段。(S410) is a means for judging pass/fail by comparing the color data Xp with the measured color data Xap1 and/or comparing the target color with the color of the plate coated with the actual candidate paint CMap1.
(S411)是当在所述手段S410中不合格时，达到合格为止将所述手段S405～S410或S407～S410反复进行的手段。(S411) is a means for repeatedly performing the above-mentioned means S405 to S410 or S407 to S410 until the above-mentioned means S410 fails to pass.
在此，“1种以上的组合物”、“登记有1种以上的组合物C1～Cn(n为2以上的整数)的各自配合组成数据Y1～Yn和分别对应于各配合组成数据的色彩数据X1～Xn的数据库”及“利用登记在该数据库中的数据的配色计算逻辑发挥作用的计算机”实质上可以与所述本发明的第1实施形态所涉及的在基于计算机调色的涂料的制作方法中被使用的装置中的“1种以上的组合物”、“数据库”及“计算机”相同。Here, "one or more compositions", "a database in which the respective combination composition data Y1 to Yn of one or more compositions C1 to Cn (n is an integer greater than 2) and the color data X1 to Xn corresponding to each combination composition data are registered" and "a computer that operates using the color matching calculation logic of the data registered in the database" can essentially be the same as the "one or more compositions", "database" and "computer" in the device used in the method for preparing a paint based on computer color matching involved in the first embodiment of the present invention.
另外，由于所述手段S401～S411相当于本发明的第1实施形态所涉及的基于计算机调色的涂料的制作方法中的用于实施S101～S111工序的手段，因此实质上可以是与关于S101～S111工序进行说明的手段相同的手段。而且，即使关于自动调和手段，也可以与通过本发明的第1实施形态所涉及的在基于计算机调色的涂料的制作方法中被使用的自动调和机来实现的自动调和手段相同。In addition, since the means S401 to S411 are equivalent to the means for implementing the steps S101 to S111 in the method for producing a paint by computer color matching according to the first embodiment of the present invention, they can be substantially the same means as the means described with respect to the steps S101 to S111. Furthermore, the automatic blending means can be the same as the automatic blending means implemented by the automatic blending machine used in the method for producing a paint by computer color matching according to the first embodiment of the present invention.
计算机调色系统(本发明的第5实施形态)Computer color matching system (fifth embodiment of the present invention)
本发明的第5实施形态所涉及的计算机调色系统是具备数据库和计算机的计算机调色系统且包括下述手段S501～S511，在该数据库中登记有1种以上的组合物的色彩数据X及配合组成数据Y，在该计算机中利用登记在该数据库中的数据的配色计算逻辑发挥作用。The computer color matching system involved in the fifth embodiment of the present invention is a computer color matching system having a database and a computer and includes the following means S501 to S511, in which color data X and matching composition data Y of one or more combinations of compositions are registered in the database, and in which a color matching calculation logic using the data registered in the database is functioned.
(S501)是使用登记在所述数据库中的数据，向所述计算机输入学习用数据的手段。(S501) is a means for inputting learning data into the computer using the data registered in the database.
(S502)是使所述学习用数据进行机器学习，生成包含从配合组成数据Y推断色彩数据X的人工智能模型的至少1种的已学习的人工智能模型的手段。(S502) is a means for subjecting the learning data to machine learning to generate at least one learned artificial intelligence model including an artificial intelligence model for inferring color data X from combination composition data Y.
(S503)是取得作为目标的色彩的目标色彩数据Xt的手段。(S503) is a means for acquiring target color dataXt of a target color.
(S504)是向所述计算机输入所述目标色彩数据Xt的手段。(S504) is a means for inputting the target color dataXt into the computer.
(S505)是通过使用计算机的检索，取得近似于所述目标色彩数据Xt的检索色彩数据Xn1及对应于检索色彩数据Xn1的近似配合组成数据Yn1，同时对所述目标色彩数据Xt与所述检索色彩数据Xn1进行比较而判定合格与否的手段。(S505) is a means for obtaining search color dataXn1 similar to the target color dataXt and approximate matching composition dataYn1 corresponding to the search color dataXn1 by using a computer search, and comparing the target color dataXt with the search color dataXn1 to determine whether they are qualified or not.
(S506)是当在所述手段S505中不合格时，使用计算机取得预测为提供目标色彩数据Xt的候补配合组成数据Yni，之后使用所述至少1种已学习的人工智能模型及/或所述人工智能模型以外的预测式，取得从候补配合组成数据Yni预测的预测色彩数据Xni，同时对所述色彩数据Xt与所述预测色彩数据Xni进行比较而判定合格与否的手段。(S506) is a method for obtaining candidate combination composition data Yni predicted to provide target color data Xt using a computer when the method S505 fails, and then obtaining predicted color data Xni predicted from the candidate combination composition data Yni using the at least one learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model, and comparing the color data Xt with the predicted color data Xni to determine whether the color data X t is qualified or not.
(S507)是当在所述手段S506中不合格时，使用计算机取得预测为提供目标色彩数据Xt的候补配合组成数据Yni，之后使用所述至少1种已学习的人工智能模型及/或所述人工智能模型以外的预测式，取得从候补配合组成数据Yni预测的预测色彩数据Xni，同时对所述色彩数据Xt与所述预测色彩数据Xni进行比较，达到合格为止将判定合格与否的手段反复进行的手段。(S507) is a method of, when the method S506 fails, using a computer to obtain candidate combination composition data Yni predicted to provide target color data Xt , then using the at least one learned artificial intelligence model and/or a prediction formula other than the artificial intelligence model to obtain predicted color data Xni predicted from the candidate combination composition data Yni , and at the same time comparing the color data Xt with the predicted color data Xni , and repeatedly performing the method of judging whether the color data X t is qualified until the color data X t is qualified.
(S508)是当在所述手段S505～S507的任意一个中合格时，取得合格配合组成数据YC1的手段。(S508) is a means for acquiring qualified combination composition dataYC1 when any one of the means S505 to S507 is qualified.
(S509)是根据所述合格配合组成数据YC1调制实际候补涂料CMCi，得到该实际候补涂料CMCi的涂装板而取得实测色彩数据XCi的手段。(S509) is a means for modulating the actual candidate paint CMCi based on the acceptable blend composition data YC1 , obtaining a coating plate of the actual candidate paint CMCi , and acquiring the measured color data XCi .
(S510)是通过所述色彩数据Xt与所述实测色彩数据XCi的比较及/或作为所述目标的色彩与所述实际候补涂料CMCi的涂装板的色彩的比较，判定合格与否的手段。(S510) is a means for judging pass/fail by comparing the color dataXt with the measured color dataXCi and/or comparing the target color with the color of the plate coated with the actual candidate paintCMCi .
(S511)是当在所述手段S510中不合格时，将所述手段S506～S510反复进行的手段。(S511) is a means of repeatedly performing the means S506 to S510 when the means S510 fails.
在此，“1种以上的组合物”、“登记有1种以上的组合物的色彩数据X及配合组成数据Y的数据库”及“利用登记在该数据库中的数据的配色计算逻辑发挥作用的计算机”实质上可以与所述本发明的第2实施形态所涉及的在基于计算机调色的涂料的制作方法中被使用的装置中的“1种以上的组合物”、“数据库”及“计算机”相同。Here, "one or more compositions", "a database in which color data X and matching composition data Y of one or more compositions are registered", and "a computer that operates using the color matching calculation logic of the data registered in the database" can be substantially the same as the "one or more compositions", "database", and "computer" in the device used in the method for preparing a paint based on computer color matching involved in the second embodiment of the present invention.
另外，由于所述手段S501～S511相当于本发明的第2实施形态所涉及的基于计算机调色的涂料的制作方法中的用于实施S201～S211工序的手段，因此实质上可以是与关于S201～S211工序进行说明的手段相同的手段。而且，即使关于自动调和手段，也可以与通过本发明的第2实施形态所涉及的在基于计算机调色的涂料的制作方法中被使用的自动调和机来实现的自动调和手段相同。In addition, since the means S501 to S511 are equivalent to the means for implementing the steps S201 to S211 in the method for producing a paint by computer color matching according to the second embodiment of the present invention, they can be substantially the same means as the means described with respect to the steps S201 to S211. Moreover, the automatic blending means can be the same as the automatic blending means implemented by the automatic blending machine used in the method for producing a paint by computer color matching according to the second embodiment of the present invention.
预测涂膜的色彩数据的系统(本发明的第6实施形态)System for predicting color data of coating film (sixth embodiment of the present invention)
本发明的第6实施形态所涉及的预测涂膜的色彩数据的系统是具备数据库和计算机的预测涂膜的色彩数据的系统且包括下述手段S601～S607，在该数据库中登记有1种以上的组合物的色彩数据X及配合组成数据Y，在该计算机中利用登记在该数据库中的数据的配色计算逻辑发挥作用。The system for predicting the color data of a coating film involved in the sixth embodiment of the present invention is a system for predicting the color data of a coating film having a database and a computer and includes the following means S601 to S607, in which color data X and combination composition data Y of one or more compositions are registered in the database, and in which a color matching calculation logic of the data registered in the database is used.
(S601)是使用登记在所述数据库中的数据，向所述计算机输入学习用数据的手段。(S601) is a means for inputting learning data into the computer using the data registered in the database.
(S602)是使所述学习用数据进行机器学习，生成包含从配合组成数据Y推断色彩数据X的人工智能模型的至少1种的已学习的人工智能模型的手段。(S602) is a means for subjecting the learning data to machine learning to generate at least one learned artificial intelligence model including an artificial intelligence model for inferring color data X from combination composition data Y.
(S603)是取得预测涂膜的色彩数据的涂料CMt的配合组成数据YCM的手段。(S603) is a means for acquiring the mixing composition data YCM of the paint CMt which is the color data of the predicted coating film.
(S604)是向所述计算机输入所述配合组成数据YCM的手段。(S604) is a means for inputting the coordination composition data YCM into the computer.
(S605)是根据需要，通过使用计算机的检索，取得对应于所述配合组成数据YCM的检索色彩数据Xn1的手段。(S605) is a means for acquiring search color dataXn1 corresponding to the coordinated composition dataYCM by searching using a computer as needed.
(S606)是当在所述手段S605中未检索到对应的检索色彩数据Xn1时，或者当并未进行所述手段S605时，使用所述至少1种已学习的人工智能模型或所述至少1种已学习的人工智能模型和所述人工智能模型以外的预测式，从所述配合组成数据YCM取得预测色彩数据Xm1的手段。(S606) is a means for obtaining predicted color dataXm1 from the combination composition data YCM using the at least one learned artificial intelligence model or the at least one learned artificial intelligence model and a prediction formula other than the artificial intelligence model when the corresponding search color dataXn1 is not retrieved in the meansS605 or when the means S605 is not performed.
(S607)是根据需要，取得涂装有所述涂料CMt的涂装板的实测色彩数据XCM，与所述预测色彩数据Xm1进行比较的手段。(S607) is a means for obtaining the measured color data XCM of the painted plate painted with the paint CMt as needed, and comparing it with the predicted color data Xm1 .
在此，“1种以上的组合物”、“登记有1种以上的组合物的色彩数据X及配合组成数据Y的数据库”及“利用登记在该数据库中的数据的配色计算逻辑发挥作用的计算机”实质上可以与本发明的第3实施形态所涉及的预测涂膜的色彩数据的方法中被使用的装置中的“1种以上的组合物”、“数据库”及“计算机”相同。Here, "one or more compositions", "a database in which color data X and matching composition data Y of one or more compositions are registered", and "a computer that operates using the color matching calculation logic of the data registered in the database" can be substantially the same as the "one or more compositions", "database", and "computer" in the device used in the method for predicting the color data of the coating film involved in the third embodiment of the present invention.
另外，由于所述手段S601～S607相当于本发明的第3实施形态所涉及的预测涂膜的色彩数据的方法中的用于实施S301～S307工序的手段，因此实质上可以是与关于S301～S307工序进行说明的手段相同的手段。而且，即使关于自动调和手段，也可以与通过本发明的第3实施形态所涉及的预测涂膜的色彩数据的方法中被使用的自动调和机来实现的自动调和手段相同。In addition, since the means S601 to S607 are equivalent to the means for implementing the steps S301 to S307 in the method for predicting the color data of the coating film according to the third embodiment of the present invention, they can be substantially the same means as the means described with respect to the steps S301 to S307. Furthermore, the automatic blending means can be the same as the automatic blending means implemented by the automatic blending machine used in the method for predicting the color data of the coating film according to the third embodiment of the present invention.
应用程序软件(本发明的第7实施形态)Application software (seventh embodiment of the present invention)
本发明还涉及一种用于对所述本发明的第4及第5实施形态所涉及的计算机调色系统以及/或所述本发明的第6实施形态所涉及的预测涂膜的色彩数据的系统进行控制并使其工作的应用程序软件。The present invention also relates to application software for controlling and operating the computer color matching system according to the fourth and fifth embodiments of the present invention and/or the system for predicting color data of a coating film according to the sixth embodiment of the present invention.
本发明的第7实施形态所涉及的应用程序软件是用于对本发明的第4及第5实施形态所涉及的计算机调色系统以及/或本发明的第6实施形态所涉及的预测涂膜的色彩数据的系统进行控制并使其工作而发挥功能的应用程序软件，由此发挥用于执行本发明的方法的功能。The application software involved in the 7th embodiment of the present invention is an application software used to control and operate the computer color matching system involved in the 4th and 5th embodiments of the present invention and/or the system for predicting the color data of the coating film involved in the 6th embodiment of the present invention to perform its function, thereby performing the function of executing the method of the present invention.
本发明的第7实施形态所涉及的应用程序软件还可以预先储存在HDD(Hard DiskDrive)或闪存器等记录装置中，该HDD(Hard Disk Drive)或闪存器等记录装置由执行本发明的第4及第5实施形态所涉及的计算机调色系统以及/或本发明的第6实施形态所涉及的预测涂膜的色彩数据的系统或构成系统的各手段的仪器等所具有。另外，还可以通过使用无线或有线通信手段、DVD、CD-ROM、USB存储器等可装拆的记录介质等而安装(install)于仪器等。The application software involved in the seventh embodiment of the present invention may also be pre-stored in a recording device such as a HDD (Hard Disk Drive) or a flash memory, which is owned by an instrument or the like that executes the computer color matching system involved in the fourth and fifth embodiments of the present invention and/or the system for predicting the color data of the coating film involved in the sixth embodiment of the present invention or each means constituting the system. In addition, the application software may be installed in an instrument or the like by using a removable recording medium such as a wireless or wired communication means, a DVD, a CD-ROM, or a USB memory.
实施例Example
以下，通过实施例更加具体地对本发明进行说明。但是，本发明并不限定于以下的实施例。Hereinafter, the present invention will be described in more detail by way of examples, but the present invention is not limited to the following examples.
实施例1Example 1
通过RETAN PG80、RETAN PG HYBRIDECO、RETAN WB ECO EV及RETAN ECO FLEET(均为关西涂料公司制商品名)系列，将彩色原色、金属原色及珠光原色总计选定了86种。取得该配合组成数据及色彩数据并登记在数据库中。色彩数据使用了用多角度分光光度计(入射角度45度、受光角度强光15度、25度、正面(face)45度、遮光面(shade)75度、110度)进行测定的5个角度400～700nm下的155种反射光谱数据。将使用登记在数据库中的数据制作的学习用数据，输入到计算机而通过神经网络进行机器学习，生成了包含从色彩数据X推断配合组成数据Y的人工智能模型的至少1种的已学习的人工智能模型。A total of 86 primary colors, metallic primary colors, and pearlescent primary colors were selected from the RETAN PG80, RETAN PG HYBRIDECO, RETAN WB ECO EV, and RETAN ECO FLEET series (all trade names of Kansai Paint Co., Ltd.). The blending composition data and color data were obtained and registered in the database. The color data used 155 types of reflectance spectrum data at 400 to 700 nm at five angles measured using a multi-angle spectrophotometer (incident angle 45 degrees, light receiving angle strong light 15 degrees, 25 degrees, front (face) 45 degrees, shaded surface (shade) 75 degrees, 110 degrees). The learning data created using the data registered in the database was input into a computer and machine learning was performed using a neural network, generating at least one learned artificial intelligence model including an artificial intelligence model that infers blending composition data Y from color data X.
准备涂布有配合组成数据不明的涂料的涂装板100张(色彩、涂膜构成不同的100个色，金属及/或珠光涂膜含有约14％)，在取得各自的色彩数据的同时，输入到同时搭载有所述已学习的人工智能模型和人工智能模型以外的预测式的计算机色彩校正装置(关西涂料公司制)，使用所述已学习的人工智能模型进行调色作业。通过调色精度及削减效果对调色作业时的调色负担进行了评价。求出调色次数在2次以内时合格的比例，以下述1～5阶段评价了调色精度。另外，削减效果则用当比较例1作为100时的比例表示了取得最终合格调色配合为止所需的工时(时间)。表1中表示结果。调色作业时的调色负担为，当比较例1作为100时，实施例1为40，从现有的调色工时有60％的削减效果。100 painted plates coated with paint whose composition data is unknown (100 colors with different colors and coating film compositions, containing about 14% of metallic and/or pearlescent coating films) were prepared, and while obtaining the respective color data, they were input into a computer color correction device (manufactured by Kansai Paint Co., Ltd.) equipped with both the learned artificial intelligence model and a predictive type other than the artificial intelligence model, and the learned artificial intelligence model was used to perform the color matching operation. The color matching burden during the color matching operation was evaluated by the color matching accuracy and reduction effect. The proportion of qualified colors within 2 times was calculated, and the color matching accuracy was evaluated in the following 1 to 5 stages. In addition, the reduction effect is expressed as the ratio when Comparative Example 1 is taken as 100, which represents the working hours (time) required to obtain the final qualified color matching. The results are shown in Table 1. The color matching burden during the color matching operation is 40 for Example 1 when Comparative Example 1 is taken as 100, and there is a 60% reduction effect from the existing color matching hours.
5：合格率为85％以上5: The pass rate is more than 85%
4：合格率为75％以上、小于85％4: The pass rate is above 75% and less than 85%
3：合格率为65％以上、小于75％3: The pass rate is above 65% and less than 75%
2：合格率为55％以上、小于65％2: The pass rate is 55% or more and less than 65%
1：合格率小于55％1: The pass rate is less than 55%
比较例1Comparative Example 1
将与实施例1相同的彩色原色、金属原色及珠光原色总计取得86种配合组成数据及色彩数据，登记在并未搭载有人工智能模型的现有型号的计算机色彩校正装置(关西涂料公司制)。A total of 86 kinds of combination composition data and color data of the same chromatic primary colors, metallic primary colors and pearlescent primary colors as in Example 1 were obtained and registered in a conventional computer color calibration device (manufactured by Kansai Paint Co., Ltd.) that was not equipped with an artificial intelligence model.
对于与实施例1相同的100张涂装板，使用上述计算机色彩校正装置得到候补配合组成，5年以上的熟练者取得合格配合为止反复进行调色作业，与实施例1同样地评价了调色负担。表1中表示结果。For the same 100 painted plates as in Example 1, candidate blending compositions were obtained using the above-mentioned computer color calibration device, and the coloring operation was repeated until a skilled person with 5 years of experience or more obtained a qualified blend, and the coloring burden was evaluated in the same manner as in Example 1. Table 1 shows the results.
实施例2Example 2
通过RETAN PG80、RETAN PG HYBRIDECO、RETAN WB ECO EV及RETAN ECO FLEET(均为关西涂料公司制商品名)系列，将彩色原色及珠光原色总计选定了86种。取得该配合组成数据及色彩数据并登记在数据库中。色彩数据使用了用多角度分光光度计(入射角度45度、受光角度强光15度、25度、正面(face)45度、遮光面(shade)75度、110度)进行测定的5个角度400～700nm下的155种反射光谱数据。将使用登记在数据库中的数据制作的学习用数据，输入到计算机而通过神经网络进行机器学习，生成了从配合组成数据Y推断色彩数据X的人工智能模型。A total of 86 primary colors and pearlescent primary colors were selected from the RETAN PG80, RETAN PG HYBRIDECO, RETAN WB ECO EV and RETAN ECO FLEET series (all product names of Kansai Paint Co., Ltd.). The blending composition data and color data were obtained and registered in the database. The color data used 155 types of reflectance spectrum data at 400 to 700 nm at five angles measured using a multi-angle spectrophotometer (incident angle 45 degrees, light receiving angle strong light 15 degrees, 25 degrees, front (face) 45 degrees, shaded surface (shade) 75 degrees, 110 degrees). The learning data created using the data registered in the database was input into a computer and machine learning was performed using a neural network to generate an artificial intelligence model that infers color data X from blending composition data Y.
准备涂布有配合组成数据不明的涂料的涂装板100张(色彩、涂膜构成不同的100个色，金属及/或珠光涂膜含有约14％)，在取得各自的色彩数据的同时，输入到同时搭载有已学习的人工智能模型和人工智能模型以外的预测式的计算机色彩校正装置(关西涂料公司制)，使用人工智能模型进行调色作业。与实施例1同样地通过调色精度及削减效果对调色作业时的调色负担进行了评价。表1中表示结果。调色作业时的调色负担为，当比较例2作为100时，实施例2为40，从现有的调色工时有60％的削减效果。100 painted plates coated with paint whose composition data is unknown (100 colors with different colors and coating film structures, containing about 14% metallic and/or pearlescent coatings) are prepared, and while obtaining the color data of each, they are input into a computer color correction device (manufactured by Kansai Paint Co., Ltd.) equipped with both a learned artificial intelligence model and a predictive model other than the artificial intelligence model, and the artificial intelligence model is used to perform color matching. The color matching burden during the color matching operation is evaluated by the color matching accuracy and reduction effect in the same manner as in Example 1. The results are shown in Table 1. The color matching burden during the color matching operation is 40 for Example 2 when Comparative Example 2 is taken as 100, which is a 60% reduction effect on the existing color matching man-hours.
比较例2Comparative Example 2
将与实施例2相同的彩色原色及珠光原色总计取得86种配合组成数据及色彩数据，登记在并未搭载有人工智能模型的现有型号的计算机色彩校正装置(关西涂料公司制)。A total of 86 kinds of blending composition data and color data were obtained for the same primary colors and pearlescent primary colors as in Example 2, and registered in a conventional computer color calibration device (manufactured by Kansai Paint Co., Ltd.) that was not equipped with an artificial intelligence model.
对于与实施例2相同的100张涂装板，使用上述计算机色彩校正装置得到候补配合组成，5年以上的熟练者取得合格配合为止反复进行调色作业，与实施例2同样地评价了调色负担。表1中表示结果。For the same 100 painted plates as in Example 2, candidate blending compositions were obtained using the above-mentioned computer color calibration device, and the coloring operation was repeated until a skilled person with 5 years of experience or more obtained a qualified blend, and the coloring burden was evaluated in the same manner as in Example 2. Table 1 shows the results.
实施例3Example 3
代替通过神经网络的机器学习，将使用学习用数据的机器学习做成使用梯度提升的决策树的机器学习，除此之外与实施例2相同，而生成了从配合组成数据Y推断色彩数据X的人工智能模型。做成与实施例2同样，使用人工智能模型进行了调色作业，通过调色精度及削减效果评价了调色作业时的调色负担。表1中表示结果。The same method as in Example 2 was used except that the machine learning using the learning data was replaced by machine learning using a decision tree using gradient boosting instead of machine learning using the neural network, and an artificial intelligence model for inferring the color data X from the combination composition data Y was generated. The same method as in Example 2 was used to perform a color adjustment operation using the artificial intelligence model, and the color adjustment burden during the color adjustment operation was evaluated by the color adjustment accuracy and reduction effect. The results are shown in Table 1.
表1Table 1
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2019-238855 | 2019-12-27 | ||
| JP2019238855AJP6703639B1 (en) | 2019-12-27 | 2019-12-27 | Paint manufacturing method and method for predicting color data |
| JP2020091802 | 2020-05-26 | ||
| JP2020-091802 | 2020-05-26 | ||
| PCT/JP2020/048972WO2021132654A1 (en) | 2019-12-27 | 2020-12-25 | Method for producing paint material, method for predicting hue data, and computer color-mixing system |
| Publication Number | Publication Date |
|---|---|
| CN115244149A CN115244149A (en) | 2022-10-25 |
| CN115244149Btrue CN115244149B (en) | 2023-09-15 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202080089917.7AActiveCN115244149B (en) | 2019-12-27 | 2020-12-25 | Paint manufacturing method, color data prediction method and computer color matching system |
| Country | Link |
|---|---|
| JP (1) | JPWO2021132654A1 (en) |
| CN (1) | CN115244149B (en) |
| WO (1) | WO2021132654A1 (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20240302214A1 (en)* | 2023-03-08 | 2024-09-12 | The EOSYS Group, Inc. | Colorimetric compliance control device |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP7116266B1 (en)* | 2020-12-14 | 2022-08-09 | 日本ペイントホールディングス株式会社 | Prediction method of paint properties, prediction method of correction composition composition, prediction system of paint properties, prediction system of correction composition correction, and paint manufacturing method |
| US20250224275A1 (en)* | 2022-03-28 | 2025-07-10 | Ppg Industries Ohio, Inc. | Techniques for color batch correction |
| JP7634284B2 (en)* | 2022-12-27 | 2025-02-21 | 日本ペイント・インダストリアルコーティングス株式会社 | Paint manufacturing method |
| JP7610317B1 (en)* | 2024-06-27 | 2025-01-08 | 日本ペイント・インダストリアルコーティングス株式会社 | Method for producing coating composition |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09160194A (en)* | 1995-12-04 | 1997-06-20 | Konica Corp | Color image forming method and formation of color proof |
| CA2459635A1 (en)* | 2003-03-03 | 2004-09-03 | Jozef Theresia Huybrechts | Two-component coating compositions |
| CN1625222A (en)* | 2003-12-02 | 2005-06-08 | 富士全录株式会社 | Image forming device, color calibration method and its program |
| CN1692147A (en)* | 2002-12-10 | 2005-11-02 | 日本油漆株式会社 | Paint decision method, paint manufacturing method, coating method, paint decision server, and paint decision program |
| CN101175979A (en)* | 2005-03-18 | 2008-05-07 | 威士伯来源有限公司 | Numerical method for matching colorants |
| CN101668109A (en)* | 2009-10-16 | 2010-03-10 | 浙江理工大学 | Soft-proofing method for color-spinning yarn by tri-stimulus value matching |
| CN103153487A (en)* | 2010-06-25 | 2013-06-12 | 纳幕尔杜邦公司 | System for producing and delivering matching color coating and use thereof |
| CN103874770A (en)* | 2011-08-08 | 2014-06-18 | 卡里斯生命科学卢森堡控股有限责任公司 | Biomarker compositions and methods |
| CN106455979A (en)* | 2013-12-31 | 2017-02-22 | 纪念斯隆-凯特琳癌症中心 | Systems, methods and devices for real-time multi-channel imaging of fluorescent sources |
| CN107330454A (en)* | 2017-06-20 | 2017-11-07 | 西安建筑科技大学 | The non-linear visualization of magnanimity higher-dimension sequence data sort feature and quantitative analysis method |
| CN107993373A (en)* | 2015-08-03 | 2018-05-04 | 天使游戏纸牌股份有限公司 | Check the check device of valuables |
| CN108139271A (en)* | 2015-10-29 | 2018-06-08 | 巴斯夫涂料有限公司 | The method for determining the parametric texture of face coat |
| CN108690416A (en)* | 2017-03-29 | 2018-10-23 | 宋承原 | Macromolecule-ceramic hybrid coating composition and the accumulator separation film manufacturing method for using above-mentioned composition |
| CN108824767A (en)* | 2018-06-22 | 2018-11-16 | 中电建建筑集团有限公司 | A kind of flame-proof environmental protection outer thermal insulation construction method and system, the information processing terminal |
| CN110046635A (en)* | 2017-12-06 | 2019-07-23 | 涂层国外知识产权有限公司 | For matching the color of target coating and the system and method for appearance |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2554014B2 (en)* | 1993-11-15 | 1996-11-13 | 日本ペイント株式会社 | Computer color matching method |
| CA2384127A1 (en)* | 2000-07-07 | 2002-01-17 | Kansai Paint Co., Ltd. | Toning method of paint having brilliant feeling |
| US6804390B2 (en)* | 2001-02-07 | 2004-10-12 | Basf Corporation | Computer-implemented neural network color matching formulation applications |
| JP2004224966A (en)* | 2003-01-24 | 2004-08-12 | Kansai Paint Co Ltd | Method for toning metallic coating color |
| US20060031027A1 (en)* | 2004-08-03 | 2006-02-09 | Alman David H | Method and apparatus for predicting properties of a chemical mixture |
| WO2008156147A1 (en)* | 2007-06-20 | 2008-12-24 | Kansai Paint Co., Ltd. | Coating color database creating method, search method using the database, their system, program, and recording medium |
| JP2009079227A (en)* | 2008-10-31 | 2009-04-16 | Nippon Paint Co Ltd | Computer toning method for paint liquid, method for producing paint using this method, and toning apparatus for paint liquid |
| MX2013015083A (en)* | 2011-06-20 | 2014-05-14 | Coatings Foreign Ip Co Llc | METHOD FOR EQUALIZING THE APPEARANCE OF THE GLOSSY OF COATINGS. |
| CN105653701B (en)* | 2015-12-31 | 2019-01-15 | 百度在线网络技术（北京）有限公司 | Model generating method and device, word assign power method and device |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09160194A (en)* | 1995-12-04 | 1997-06-20 | Konica Corp | Color image forming method and formation of color proof |
| CN1692147A (en)* | 2002-12-10 | 2005-11-02 | 日本油漆株式会社 | Paint decision method, paint manufacturing method, coating method, paint decision server, and paint decision program |
| CA2459635A1 (en)* | 2003-03-03 | 2004-09-03 | Jozef Theresia Huybrechts | Two-component coating compositions |
| CN1625222A (en)* | 2003-12-02 | 2005-06-08 | 富士全录株式会社 | Image forming device, color calibration method and its program |
| CN101175979A (en)* | 2005-03-18 | 2008-05-07 | 威士伯来源有限公司 | Numerical method for matching colorants |
| CN101668109A (en)* | 2009-10-16 | 2010-03-10 | 浙江理工大学 | Soft-proofing method for color-spinning yarn by tri-stimulus value matching |
| CN103153487A (en)* | 2010-06-25 | 2013-06-12 | 纳幕尔杜邦公司 | System for producing and delivering matching color coating and use thereof |
| CN103874770A (en)* | 2011-08-08 | 2014-06-18 | 卡里斯生命科学卢森堡控股有限责任公司 | Biomarker compositions and methods |
| CN106455979A (en)* | 2013-12-31 | 2017-02-22 | 纪念斯隆-凯特琳癌症中心 | Systems, methods and devices for real-time multi-channel imaging of fluorescent sources |
| CN107993373A (en)* | 2015-08-03 | 2018-05-04 | 天使游戏纸牌股份有限公司 | Check the check device of valuables |
| CN108139271A (en)* | 2015-10-29 | 2018-06-08 | 巴斯夫涂料有限公司 | The method for determining the parametric texture of face coat |
| CN108690416A (en)* | 2017-03-29 | 2018-10-23 | 宋承原 | Macromolecule-ceramic hybrid coating composition and the accumulator separation film manufacturing method for using above-mentioned composition |
| CN107330454A (en)* | 2017-06-20 | 2017-11-07 | 西安建筑科技大学 | The non-linear visualization of magnanimity higher-dimension sequence data sort feature and quantitative analysis method |
| CN110046635A (en)* | 2017-12-06 | 2019-07-23 | 涂层国外知识产权有限公司 | For matching the color of target coating and the system and method for appearance |
| CN108824767A (en)* | 2018-06-22 | 2018-11-16 | 中电建建筑集团有限公司 | A kind of flame-proof environmental protection outer thermal insulation construction method and system, the information processing terminal |
| Title |
|---|
| Combinatorial screening for new materials in unconstrained composition space with machine learning;B. Meredig等;《PHYSICAL REVIEW B》;第89卷;第094104（1-7）页* |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20240302214A1 (en)* | 2023-03-08 | 2024-09-12 | The EOSYS Group, Inc. | Colorimetric compliance control device |
| Publication number | Publication date |
|---|---|
| CN115244149A (en) | 2022-10-25 |
| JPWO2021132654A1 (en) | 2021-07-01 |
| WO2021132654A1 (en) | 2021-07-01 |
| Publication | Publication Date | Title |
|---|---|---|
| JP6703639B1 (en) | Paint manufacturing method and method for predicting color data | |
| JP6936416B1 (en) | How to make paint and how to predict color data | |
| CN115244149B (en) | Paint manufacturing method, color data prediction method and computer color matching system | |
| CN108139271B (en) | Method for determining texture parameters of surface coatings | |
| EP2161555B1 (en) | Coating color database creating method, search method using the database, their system, program, and recording medium | |
| US6714924B1 (en) | Computer-implemented neural network color matching formulation system | |
| JP5306519B2 (en) | IDENTIFICATION METHOD, IDENTIFICATION SYSTEM, IDENTIFICATION PROGRAM, AND RECORDING MEDIUM THEREOF | |
| US6804390B2 (en) | Computer-implemented neural network color matching formulation applications | |
| KR101162078B1 (en) | Method for matching paint | |
| JP5063076B2 (en) | IDENTIFICATION METHOD, IDENTIFICATION SYSTEM, IDENTIFICATION PROGRAM, AND RECORDING MEDIUM THEREOF | |
| CN113128548A (en) | System and method for matching color and appearance of target coating | |
| US20030208345A1 (en) | Color matching and simulation of multicolor surfaces | |
| CN108875792A (en) | A kind of coating based on machine learning repairs color method and system | |
| CN113128547A (en) | System and method for matching color and appearance of target coating | |
| US20250237638A1 (en) | Method and system for determining color match for surface coatings | |
| JP3870421B2 (en) | Computer color matching method and apparatus | |
| US20040239928A1 (en) | Modified-color generation and display method and apparatus | |
| US20140350867A1 (en) | System for producing liquid composition | |
| JP7417789B2 (en) | Fully integrated digital color management system | |
| JP2003294530A (en) | Computer color matching calculation method | |
| Romney | Relating reflectance spectra space to Munsell color appearance space | |
| Li et al. | Spectral prediction method based on the transformer neural network for high-fidelity color reproduction | |
| JP7681784B1 (en) | Method for predicting changes in coating film, system for predicting changes in coating film, and application software | |
| JP2003034762A (en) | Method for toning of coating having lustrous shining | |
| JP2006030210A (en) | Method and system of computer color matching |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |