技术领域technical field
本发明涉及图像缺陷识别领域，特别涉及一种应用于手机外壳复杂纹理背景的缺陷检测方法。The invention relates to the field of image defect recognition, in particular to a defect detection method applied to a complex texture background of a mobile phone shell.
背景技术Background technique
本发明方法涉及到计算机视觉的图像处理，特别是图像识别这一部分，从一张放大的手机外壳图片中提取出缺陷位置的过程。The method of the invention relates to the image processing of computer vision, especially the part of image recognition, which is the process of extracting the defect position from an enlarged mobile phone casing picture.
与本发明最相近的方法有宋迪[1]等人针对手机配件划痕提出了一种基于Gabor和纹理抑制的手机配件划痕检测算法，对金属表面图像进行Gabor滤波,提取出划痕的骨架结构,利用各向异性纹理抑制方法抑制金属表面的纹理,再用滞后阈值准确提取划痕。The method closest to the present invention has Song Di [1] et al. propose a kind of mobile phone accessory scratch detection algorithm based on Gabor and texture suppression for the scratches of mobile phone accessories, carry out Gabor filter to the metal surface image, extract the scratches Skeleton structure, using the anisotropic texture suppression method to suppress the texture of the metal surface, and then using the hysteresis threshold to accurately extract scratches.
引用文献：Citation:
[1]宋迪,张东波,刘霞.基于Gabor和纹理抑制的手机配件划痕检测[J].计算机工程.2014,40(9):1-5.[1] Song Di, Zhang Dongbo, Liu Xia. Scratch detection of mobile phone accessories based on Gabor and texture suppression [J]. Computer Engineering. 2014,40(9):1-5.
[2]RUDIN,L.,OSHER,S.,AND FATEMI,E.1992.Nonlinear total variationbased noise removal algorithms.Physica D:Nonlinear Phenomena 60,1-4,259–268.[2]RUDIN,L.,OSHER,S.,AND FATEMI,E.1992.Nonlinear total variation based noise removal algorithms.Physica D:Nonlinear Phenomena 60,1-4,259–268.
[3]Li Xu,Qiong Yan,Yang Xia,Jiaya Jia.2012.Structure Extraction fromTexture via Relative Total Variation.ACM Transactions on Graphics,31(6),139:1-10.[3] Li Xu, Qiong Yan, Yang Xia, Jiaya Jia. 2012. Structure Extraction from Texture via Relative Total Variation. ACM Transactions on Graphics, 31(6), 139:1-10.
[4]Xianghua Xie.A review of recent advances in surface defectdetection using texture analysis techniques[J].Electronic Letters on ComputerVision and Image Analysis,2008,7(3):1-22.[4]Xianghua Xie.A review of recent advances in surface defect detection using texture analysis techniques[J].Electronic Letters on ComputerVision and Image Analysis,2008,7(3):1-22.
发明内容Contents of the invention
目前大量的算法都是在一些背景纹理比较清晰的图像中，通过颜色信息、边缘检测、Gabor滤波、小波变换等提取出图像表面比较明显的缺陷。这些方法存在的最大问题是图像背景对缺陷的干扰影响很大，如果背景纹理稍微复杂点，就难以检测出缺陷或者误检。针对这个问题，本发明针对手机外壳进行放大后的有着复杂纹理的图像，首先从复杂纹理中提取出图像的主结构，然后进行自适应的边缘检测，最后通过一系列的滤波和数学形态学运算操作，解决了这一问题。At present, a large number of algorithms extract obvious defects on the image surface through color information, edge detection, Gabor filter, wavelet transform, etc. in images with clear background textures. The biggest problem with these methods is that the image background has a great influence on the interference of defects. If the background texture is slightly more complex, it is difficult to detect defects or false detections. To solve this problem, the present invention aims at the enlarged image with complex texture of the mobile phone shell, first extracts the main structure of the image from the complex texture, then performs adaptive edge detection, and finally through a series of filtering and mathematical morphology operations operation, which solved the problem.
本发明提出一种基于手机外壳复杂纹理的缺陷检测方法，首先对手机外壳表面划分为多个部分，进行放大的预处理；然后对放大后的手机外壳图片进行图像的主结构提取，减弱放大后纹理对检测的影响；再根据一种自适应边缘检测Canny方法来进行边缘检测；最后根据一系列的中值滤波、数学形态学运算、8联通区域和面积筛选处理后提取缺陷。The invention proposes a defect detection method based on the complex texture of the mobile phone shell. Firstly, the surface of the mobile phone shell is divided into multiple parts, and the enlarged preprocessing is performed; The impact of texture on detection; then edge detection is performed according to an adaptive edge detection Canny method; finally, defects are extracted after a series of median filtering, mathematical morphological operations, and 8-connected regions and area screening.
进一步的，所述自适应边缘检测Canny方法包括以下步骤：Further, the adaptive edge detection Canny method comprises the following steps:
用高斯滤波器平滑图像；Smooth the image with a Gaussian filter;
用一阶偏导的有限差分来计算梯度的幅值和方向；Calculate the magnitude and direction of the gradient using the finite difference of the first partial derivative;
对梯度幅值应用非极大值抑制；apply non-maximum suppression to the gradient magnitude;
用双阙值算法检测和链接边缘。Edges are detected and linked using a dual-threshold algorithm.
进一步的，手机外壳缺陷分为四种，分别是擦痕、划痕、缺损、污渍。Furthermore, there are four types of mobile phone shell defects, namely scratches, scratches, defects, and stains.
附图说明Description of drawings
图1为显著性检测方法流程框图；Fig. 1 is a block diagram of the flow chart of the saliency detection method;
图2A为手机样本图；Figure 2A is a sample diagram of a mobile phone;
图2B为手机样本放大后的图；Figure 2B is an enlarged view of the mobile phone sample;
图3A为纹理原图；Figure 3A is the original texture image;
图3B为纹理抑制结果图；Fig. 3B is a result map of texture suppression;
图4为边缘检测结果图；Fig. 4 is an edge detection result figure;
图5A为中值滤波后图；Fig. 5A is a figure after median filtering;
图5B为数学形态学运算后图；Fig. 5B is the figure after mathematical morphology operation;
图6为最终结果图；Figure 6 is the final result figure;
图7A为擦痕图；Figure 7A is a scratch map;
图7B为划痕图；Figure 7B is a scratch map;
图7C为缺损图；Figure 7C is a defect map;
图7D为污渍图；Figure 7D is a stain map;
图8A为擦痕结果图；Figure 8A is a graph of scratches results;
图8B为划痕结果图；Figure 8B is a graph of scratch results;
图8C为缺损结果图；Figure 8C is a defect result map;
图8D为污渍结果图。Figure 8D is a graph of stain results.
具体实施方式detailed description
以下结合附图对本发明进行详细说明：The present invention is described in detail below in conjunction with accompanying drawing:
图1是手机外壳缺陷检测方法流程框图。主要包括划分区域放大图像、提取图像主结构、边缘检测、中值滤波、数学形态学等七个方面。Fig. 1 is a block diagram of a method for detecting defects in a mobile phone casing. It mainly includes seven aspects such as dividing the area to enlarge the image, extracting the main structure of the image, edge detection, median filter, and mathematical morphology.
图像预处理：图像预处理主要分为二个部分：对手机外壳进行区域划分、放大图像。进行区域划分：我们将拍摄的手机外壳图片按照行、列进行划分为若干个区域，为了下一部分的放大图像做准备。如图2A所示。放大图像：针对手机外壳缺陷的细小等特征，对其中的一块，采用Micro Capture数码显微镜对其进行放大，如图2B所示。Image preprocessing: Image preprocessing is mainly divided into two parts: dividing the mobile phone shell into regions and enlarging the image. Divide the area: We divide the captured mobile phone case picture into several areas according to rows and columns, in preparation for the enlarged image in the next part. As shown in Figure 2A. Enlarging the image: Aiming at the small features of the defects of the mobile phone casing, one of them is enlarged with a Micro Capture digital microscope, as shown in Figure 2B.
图像纹理抑制：图像纹理抑制是建立在手机外壳复杂背景下使用的，目的是为了消除复杂纹理背景对缺陷提取的干扰。RUDIN[2]等人在1992年提出了一个基于总变分(ROF)模型的方法，已经广泛应用在图像重构、复原、去噪等，ROF模型定义如下：Image texture suppression: Image texture suppression is based on the complex background of the mobile phone shell, and the purpose is to eliminate the interference of complex texture backgrounds on defect extraction. In 1992, RUDIN [2] et al proposed a method based on the Total Variation (ROF) model, which has been widely used in image reconstruction, restoration, denoising, etc. The ROF model is defined as follows:
其中f和g分别表示输入图像和输出图源，λ是一个平衡因子，表示总变分，用于计算f和g之间的L2距离。where f and g represent the input image and output image source respectively, and λ is a balance factor, represents the total variation, Used to calculate the L2 distance between f andg .
Li Xu[3]等人对ROF模型进行改进，模型如下：Li Xu[3] and others improved the ROF model, the model is as follows:
其中I代表输入图像，p代表2D图像像素的索引，S代表输出结构图像。where I represents the input image, p represents the index of the 2D image pixel, and S represents the output structure image.
q是以p点为中心的一个正方形区域内的所有的像素点的索引，g为高斯函数。q is the index of all pixels in a square area centered on point p, and g is a Gaussian function.
经过纹理抑制后的结果前后对比分别如图3A和图3B。The before and after comparison of the results after texture suppression is shown in Figure 3A and Figure 3B respectively.
边缘检测：Canny边缘检测算子其实质是用1个准高斯函数作平滑运算，然后以带方向的一阶微分算子定位导数最大值，根据变分方法得到高斯模板导数逼近，在理论上很接近4个指数函数线性组合形成的最佳边缘算子，采用高斯函数对图像作平滑处理，因此具有较强的去噪能力。Canny算子的具体步骤为Edge detection: the Canny edge detection operator essentially uses a quasi-Gaussian function for smoothing operations, and then uses a directional first-order differential operator to locate the maximum value of the derivative, and obtains the Gaussian template derivative approximation according to the variational method, which is theoretically very It is close to the best edge operator formed by the linear combination of four exponential functions, and uses Gaussian function to smooth the image, so it has strong denoising ability. The specific steps of the Canny operator are
用高斯滤波器平滑图像；Smooth the image with a Gaussian filter;
用一阶偏导的有限差分来计算梯度的幅值和方向；Calculate the magnitude and direction of the gradient using the finite difference of the first partial derivative;
对梯度幅值应用非极大值抑制；apply non-maximum suppression to the gradient magnitude;
用双阙值算法检测和链接边缘。Edges are detected and linked with a double-threshold algorithm.
根据上述手机外壳纹理抑制后，用Canny算子检测后的边缘检测结果如图4所示。According to the texture suppression of the above mobile phone shell, the edge detection result after detection by the Canny operator is shown in Figure 4.
中值滤波和数学形态学运算：中值滤波SM(Standard Median Filter)是一种具有较少边缘模糊的非线性滤波方法，不仅能够去除或者减少随机噪声和脉冲干扰，还能较好地保留图像边缘信息。这种算法主要依赖于快读排序算法，其基本思想是在要排序的元素集合中任意选取一个元素并将它与其他元素进行比较，将所有比这个元素小的元素都放在它之前，将所有比它大的元素放在它之后；经过一次排序之后，可按该元素所在的位置分界，将集合分成2个部分；然后对剩下的2个部分重复上述过程进行排序，直到每一部分只剩下一个元素为止；当所有排序完成后，取排序后的集合中位于中间位置的元素的值(即所谓的中值)作为输出值。传统中值滤波可以定义为：Median filtering and mathematical morphology operations: Median filtering SM (Standard Median Filter) is a nonlinear filtering method with less edge blur, which can not only remove or reduce random noise and pulse interference, but also better preserve the image edge information. This algorithm mainly relies on the fast read sorting algorithm. Its basic idea is to randomly select an element in the set of elements to be sorted and compare it with other elements, and place all elements smaller than this element before it. All elements larger than it are placed after it; after sorting once, the collection can be divided into 2 parts according to the position of the element; then repeat the above process for the remaining 2 parts until each part is only Until one element is left; when all sorting is completed, take the value of the element in the middle of the sorted set (the so-called median value) as the output value. Traditional median filtering can be defined as:
g(x,y)＝med{f(xi,yj)}(i,j)∈M (5)g(x,y)=med{f(xi ,yj )}(i,j)∈M (5)
其中g(x,y)为中值滤波输出，f(xi,yj)为图像的像素(xi,yj)的灰度值，M为模板窗口。在中值滤波前先用一次膨胀操作可以把缺陷的信息更加明显，如图5A所示。图像的形态学运算是从数学形态学的集合论发展起米的，尽管它的基本运算很简单，但它们的组合可以产生很多复杂的效果。形态学图像处理就是在图像中移动一个结构元素并进行一种类似于卷积的操作。在每个像素位置，结构元素与在它下面的二值图像之间进行一种特定的逻辑运算。逻辑运算的二进制结果保存在输出图像中对应于该像素的位置上，产生的效果取决于该结构构元素的大小、内容以及逻辑运算的性质。Where g(x,y) is the median filter output, f(xi ,yj ) is the gray value of the pixel (xi ,yj ) of the image, and M is the template window. Using an expansion operation before the median filter can make the defect information more obvious, as shown in Figure 5A. The morphological operation of images is developed from the set theory of mathematical morphology. Although its basic operations are simple, their combination can produce many complex effects. Morphological image processing is to move a structural element in the image and perform an operation similar to convolution. At each pixel position, a specific logical operation is performed between the structural element and the binary image below it. The binary result of the logical operation is stored at the position corresponding to the pixel in the output image, and the effect depends on the size and content of the structural element and the nature of the logical operation.
(1)腐蚀操作，其功能是对图像进行腐蚀操作。腐蚀运算是消除物体所有边界点的一种过程，其结果使剩下的物体沿其周边比原物体小几个像素。图像中小的物体被除去，腐蚀对图像中去除小的无意义的物体来说是很有用的。腐蚀的定义为：(1) Corrosion operation, its function is to corrode the image. Erosion is the process of eliminating all boundary points of an object, resulting in the remaining object being a few pixels smaller than the original object along its perimeter. Small objects in the image are removed, erosion is useful for removing small meaningless objects in the image. Corrosion is defined as:
其中S为结构元素，也就是说，由S对B腐蚀所产生的二值图像E是这样的点的集台，如果s的原点平移到点(x，y)那么s将完全包含于B中，点(x，y)的值就是1。where S is a structural element, that is, the binary image E produced by the erosion of B by S is a set of such points, if the origin of s is translated to point (x, y) then s will be completely contained in B , the value of point (x, y) is 1.
(2)膨胀操作，其功能是对图像进行膨胀操作。膨胀是腐蚀的反操作，是将与物体接触的所有背景点合并到该物体中的过程。结果是使物体的边界增大了。膨胀在填补分割后的物体中的空洞很有用。膨胀的定义为：(2) Expansion operation, its function is to perform expansion operation on the image. Dilation is the inverse of erosion, the process of merging all background points that are in contact with an object into that object. The result is that the bounds of the object are enlarged. Dilation is useful for filling holes in a segmented object. Expansion is defined as:
也就是说，s对B膨胀产生的二值图像D是由这样的点组成的集台，如果s的原点位移到(x，y)那么它与B的交集为非空，点(x，y)的值就是1。That is to say, the binary image D generated by the expansion of s to B is a collection of such points. If the origin of s is displaced to (x, y), then its intersection with B is non-empty, and the point (x, y ) is 1.
图像的开运算是对图像先腐蚀后膨胀的过程。它具有消除细小物体、在纤细点处分离物体和平滑较大物体的边界时不明显改变其面积的作用。闭运是图像先膨胀后腐蚀运算的结果。闭运算具有过滤功能，可填平图像内部小沟、孔洞和裂缝，使断线相连。对二值化处理后的图像先进行灰度形态学闭运算然后再进行灰度形态学开运算，有效地对缺陷的连通区域进行整合，最后再进行5*5模板的中值滤波后可以对图像进行有效的消噪，并最后进行内部填充，效果如图5B所示。The image opening operation is a process of first corroding and then dilating the image. It has the effect of eliminating small objects, separating objects at thin points, and smoothing the boundaries of larger objects without significantly changing their area. Closure is the result of image dilation first and then erosion operation. The closing operation has a filtering function, which can fill up the small grooves, holes and cracks inside the image, and make the broken lines connect. For the binarized image, the gray-scale morphological closing operation is performed first, and then the gray-scale morphological opening operation is performed to effectively integrate the connected regions of the defect. Finally, the median filter of the 5*5 template can be used to The image is effectively denoised, and finally filled inside, the effect is shown in Figure 5B.
8联通区域和面积筛选：经过一系列的滤波和形态学运算后，为了更方便的分别对缺陷候选区域进行特征判别，首先我们得对处理后的结果中的各个分离部分进行标记。这里采用了8连通邻域(像素的邻域即像素周围的八个点)来确定某一像素所属的标签。对于缺陷来说，虽然它在图像中的面积大小、形状都是不确定，但可以明确那些过于小的区域肯定不是缺陷位置。假设图像的总面积为area，其各个候选区域的面积area(i)，(其中i＝1,2,…N)时，当其面积比例N小于(在这里设置为0.001)时，我们便可以认为此块区域并非是缺陷位置，很有可能是一些较小的图像噪声等等，这样就可以根据面积比例来删除一些虚假区域，可以实现较好的图像效果处理，如图像6所示。8 Connected area and area screening: After a series of filtering and morphological operations, in order to more conveniently distinguish the features of the defect candidate area, we first have to mark each separated part of the processed result. Here, an 8-connected neighborhood (the neighborhood of a pixel is the eight points around the pixel) is used to determine the label to which a certain pixel belongs. For a defect, although its area size and shape in the image are uncertain, it can be determined that those areas that are too small are definitely not defect locations. Assuming that the total area of the image is area, and the area of each candidate area is area(i), (where i=1,2,...N), when the area ratio N is less than (here is set to 0.001), we can think that this block area is not a defect position, it is likely to be some small image noise, etc., so that some false areas can be deleted according to the area ratio, and better Image effect processing, as shown in image 6.
本发明相对于现有技术具有如下的优点及效果(本发明创造与现有技术相比，所具有的优点和积极效果。)：Compared with the prior art, the present invention has the following advantages and effects (the present invention creates advantages and positive effects compared with the prior art.):
应用到图像识别中的缺陷检测中，对于最好的技术来说，往往会忽略了一些复杂背景和细小的缺陷，这样在手机外壳等一些金属表面不能保证其粗糙度能达标，而本发明在原有的技术基础上，首先对手机外壳进行区域划分和放大，在一些边缘检测和滤波、直方图的基础应用上加入纹理抑制，保证一些细小缺陷可以充分被检测出来，完成很好的缺陷识别。When applied to defect detection in image recognition, for the best technology, some complex backgrounds and small defects are often ignored, so that the roughness of some metal surfaces such as mobile phone casings cannot be guaranteed to meet the standard, and the present invention is based on the original On the basis of some technologies, first divide and enlarge the mobile phone shell, and add texture suppression to some basic applications of edge detection, filtering and histogram to ensure that some small defects can be fully detected and complete good defect identification.
根据手机外壳的大小和表面纹理的复杂性、粗糙度的检测，对手机外壳进行区域的划分和放大，将放大处理后手机外壳缺陷分为四种，分别是擦痕、划痕、缺损、污渍，情况分别如图7A-7D所示，最后检测的结果分别如图8A-8D。According to the size of the mobile phone case and the complexity and roughness detection of the surface texture, the area of the mobile phone case is divided and enlarged, and the defects of the mobile phone case after the enlargement process are divided into four types, namely scratches, scratches, defects, and stains , the situations are shown in Figures 7A-7D respectively, and the final detection results are shown in Figures 8A-8D respectively.
上述实施例为本发明较佳的实施方式，但本发明的实施方式并不受上述实施例的限制，其他的任何未背离本发明的精神实质与原理下所作的改变、修饰、替代、组合、简化，均应为等效的置换方式，都包含在本发明的保护范围之内。The above-mentioned embodiment is a preferred embodiment of the present invention, but the embodiment of the present invention is not limited by the above-mentioned embodiment, and any other changes, modifications, substitutions, combinations, Simplifications should be equivalent replacement methods, and all are included in the protection scope of the present invention.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610755289.7ACN106372645A (en) | 2016-08-29 | 2016-08-29 | Mobile phone housing complex texture background defect detection method |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610755289.7ACN106372645A (en) | 2016-08-29 | 2016-08-29 | Mobile phone housing complex texture background defect detection method |
| Publication Number | Publication Date |
|---|---|
| CN106372645Atrue CN106372645A (en) | 2017-02-01 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610755289.7APendingCN106372645A (en) | 2016-08-29 | 2016-08-29 | Mobile phone housing complex texture background defect detection method |
| Country | Link |
|---|---|
| CN (1) | CN106372645A (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108181315A (en)* | 2017-11-25 | 2018-06-19 | 天津大学 | A kind of biscuit damage detection apparatus and detection method based on image procossing |
| CN108280838A (en)* | 2018-01-31 | 2018-07-13 | 桂林电子科技大学 | A kind of intermediate plate tooth form defect inspection method based on edge detection |
| CN117237340A (en)* | 2023-11-10 | 2023-12-15 | 江西省中鼐科技服务有限公司 | Method and system for detecting appearance of mobile phone shell based on artificial intelligence |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102313740A (en)* | 2010-07-05 | 2012-01-11 | 汉王科技股份有限公司 | Solar panel crack detection method |
| CN103886602A (en)* | 2014-03-28 | 2014-06-25 | 重庆大学 | Radial image deflect detecting method based on veins |
| CN104637067A (en)* | 2015-03-18 | 2015-05-20 | 厦门麦克玛视电子信息技术有限公司 | Method for detecting defect of textured surface |
| CN105160652A (en)* | 2015-07-10 | 2015-12-16 | 天津大学 | Handset casing testing apparatus and method based on computer vision |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102313740A (en)* | 2010-07-05 | 2012-01-11 | 汉王科技股份有限公司 | Solar panel crack detection method |
| CN103886602A (en)* | 2014-03-28 | 2014-06-25 | 重庆大学 | Radial image deflect detecting method based on veins |
| CN104637067A (en)* | 2015-03-18 | 2015-05-20 | 厦门麦克玛视电子信息技术有限公司 | Method for detecting defect of textured surface |
| CN105160652A (en)* | 2015-07-10 | 2015-12-16 | 天津大学 | Handset casing testing apparatus and method based on computer vision |
| Title |
|---|
| XIANGHUA XIE等: ""A review of recent advances in surface defect detection using texture analysis techiques"", 《ELECTRONIC LETTERS ON COMPUTER VISION AND IMAGE ANALYSIS》* |
| 宋莹等: ""基于图像分块的边缘检测算法"", 《计算机工程》* |
| 宋迪等: ""基于Gabor和纹理抑制的手机配件划痕检测"", 《计算机工程》* |
| 张馨等: ""被动式太赫兹图像目标检测研究"", 《光学学报》* |
| 杨金柱等: ""融合GVF Snake与肤色模型的手势轮廓提取方法"", 《小型微型计算机系统》* |
| 苏俊宏等: ""圆柱型高精密零件表面缺陷检测及形貌分析"", 《激光与光电子学进展》* |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108181315A (en)* | 2017-11-25 | 2018-06-19 | 天津大学 | A kind of biscuit damage detection apparatus and detection method based on image procossing |
| CN108181315B (en)* | 2017-11-25 | 2021-02-02 | 天津大学 | A kind of biscuit damage detection device and detection method based on image processing |
| CN108280838A (en)* | 2018-01-31 | 2018-07-13 | 桂林电子科技大学 | A kind of intermediate plate tooth form defect inspection method based on edge detection |
| CN117237340A (en)* | 2023-11-10 | 2023-12-15 | 江西省中鼐科技服务有限公司 | Method and system for detecting appearance of mobile phone shell based on artificial intelligence |
| CN117237340B (en)* | 2023-11-10 | 2024-01-26 | 江西省中鼐科技服务有限公司 | Method and system for detecting appearance of mobile phone shell based on artificial intelligence |
| Publication | Publication Date | Title |
|---|---|---|
| El-Sayed et al. | New edge detection technique based on the shannon entropy in gray level images | |
| CN108830857B (en) | An Adaptive Binarization Segmentation Method for Chinese Inscriptions on Inscriptions | |
| CN112614062A (en) | Bacterial colony counting method and device and computer storage medium | |
| Song et al. | Surface defect detection method using saliency linear scanning morphology for silicon steel strip under oil pollution interference | |
| Kaur et al. | Comparative study of different edge detection techniques | |
| CN111739058B (en) | Free liquid level recognition and extraction method based on watershed algorithm of Gaussian filtering | |
| CN101655977B (en) | A Method of Removing Image Impulse Noise Based on Differential Image Detection and Multi-window Filtering | |
| CN107256413A (en) | A kind of article monitoring method and device | |
| Kumar et al. | Comparative analysis for edge detection techniques | |
| CN110473222A (en) | Image-element extracting method and device | |
| Reddy et al. | Canny scale edge detection | |
| CN106372645A (en) | Mobile phone housing complex texture background defect detection method | |
| CN105447489A (en) | Character and background adhesion noise elimination method for image OCR system | |
| CN117036259A (en) | Metal plate surface defect detection method based on deep learning | |
| Haider et al. | A hybrid method for edge continuity based on Pixel Neighbors Pattern Analysis (PNPA) for remote sensing satellite images | |
| CN106446920A (en) | Stroke width transformation method based on gradient amplitude constraint | |
| Jang et al. | Binarization of noisy gray-scale character images by thin line modeling | |
| Mustapha et al. | Crack Detection on Surfaces Using Digital Image Processing | |
| Li et al. | Multi-scale morphologic tracking approach for edge detection | |
| Ushma et al. | Object detection in image processing using edge detection techniques | |
| CN104102911A (en) | Image processing for AOI (automated optical inspection)-based bullet appearance defect detection system | |
| Barbu | Automatic edge detection solution using anisotropic diffusion-based multi-scale image analysis and fine-to-coarse tracking | |
| CN107133965A (en) | One kind is based on computer graphic image morphological image segmentation method | |
| Ravikumar et al. | Applications for green technology practices in analyzing the edge detection techniques on digital image processing by using MATLAB | |
| Othman et al. | The fusion of edge detection and mathematical morphology algorithm for shape boundary recognition |
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | Application publication date:20170201 | |
| RJ01 | Rejection of invention patent application after publication |