Disclosure of Invention
The embodiment of the invention provides a goods warehousing interaction method and a system suitable for a multi-dimensional data model, which can automatically adjust warehousing strategies according to the characteristics of different goods, improve warehousing efficiency and meet the requirements of modern high efficiency.
In a first aspect of the embodiment of the present invention, a method for interaction in warehousing of goods applicable to a multidimensional data model is provided, including:
The method comprises the steps that when a server judges that a storage requirement of a first storage terminal is received, the storage requirement is sent to a multidimensional data model, wherein the multidimensional data model is provided with a pre-configured storage strategy and an interaction method, and the storage strategy at least comprises at least one of an image extraction sub-strategy, a voice extraction sub-strategy and/or a text extraction sub-strategy;
analyzing the warehousing requirements by the multidimensional data model, determining a customized warehousing strategy, and feeding back to a corresponding first warehousing end so that the first warehousing end can extract information of a warehousing target according to the warehousing strategy;
Generating a corresponding interaction path by the multidimensional data model based on an interaction method corresponding to the warehousing requirement, wherein an interaction starting point in the interaction path corresponds to the first warehousing end, and other interaction nodes in the interaction path correspond to other warehousing management ends;
After judging that the multi-dimensional data model receives the warehouse-in data extracted by the first warehouse-in end and the interactive data fed back by the warehouse-in management end, feeding back corresponding warehouse-in requirements to the server so as to enable the server to finish warehouse-in records.
Optionally, in one possible implementation manner of the first aspect, the server sends the warehousing requirement to a multidimensional data model when judging that the warehousing requirement of the first warehousing end is received, where the multidimensional data model has a preconfigured warehousing strategy and an interaction method, and the method includes:
When judging that the warehouse-in demand of the first warehouse-in end is received, the server extracts a warehouse-in article label corresponding to the warehouse-in demand and sends the warehouse-in article label to the multi-dimensional data model;
The multidimensional data model determines corresponding warehousing strategies and interaction methods based on the warehousing article labels, and each warehousing article label has a preset warehousing strategy and interaction method; the multidimensional data model analyzes the warehousing requirements, determines customized warehousing strategies and feeds the customized warehousing strategies back to corresponding first warehousing ends, so that the first warehousing ends extract information of warehousing targets according to the warehousing strategies, and the method comprises the following steps:
The method comprises the steps that a multi-dimensional data model extracts a warehousing strategy of a warehousing article label, wherein the warehousing strategy at least comprises at least one of an image extraction sub-strategy, a voice extraction sub-strategy and/or a text extraction sub-strategy;
generating a customized information extraction template with an information extraction mark based on the type information and the number information of the sub-strategies in the warehousing strategy, and feeding back to a first warehousing end;
The first warehouse-in terminal interacts with the information extraction template based on the information extraction mark, and performs information extraction on the warehouse-in target.
Optionally, in one possible implementation manner of the first aspect, the generating, based on the type information and the number information of the sub-policies in the warehousing policy, an information extraction template with an information extraction identifier, and feeding back the information extraction template to the first warehousing terminal includes:
If the warehousing strategy is judged to comprise an image extraction sub-strategy, an image sub-model in the multidimensional data model acquires a warehousing type corresponding to the warehousing object label, and historical data of the warehousing type is analyzed and processed to obtain historical image extraction characteristics corresponding to the warehousing type;
Acquiring a feature tag corresponding to each historical image extraction feature, and screening the feature tag to obtain a target feature tag, wherein the feature tag at least comprises a view tag and a wear tag;
And extracting a target view label and a target loss label corresponding to the target feature label by the image sub-model, and establishing a corresponding image slot in the information extraction template based on the target view label and the target loss label.
Optionally, in one possible implementation manner of the first aspect, the obtaining a feature tag corresponding to each extracted feature of the historical image, and filtering the feature tag to obtain a target feature tag, where the feature tag includes at least a view tag and a wear tag includes:
Counting the abnormal quantity of abnormal wear-out labels in each view label in all historical image extraction characteristics of the warehouse-in article labels of the same kind, and counting the normal quantity of normal wear-out labels in each view label;
Generating a wear weight of each wear tag based on the description information of the wear tag, wherein each description information has a preset wear weight;
And calculating based on the abnormal quantity, the normal quantity and the wear weight to obtain feature tag coefficients of the extracted features of the corresponding historical images, and determining a target feature tag from a plurality of feature tags based on the feature tag coefficients.
Optionally, in one possible implementation manner of the first aspect, the calculating based on the abnormal number, the normal number and the wear weight, to obtain a feature tag coefficient of the extracted feature of the corresponding historical image, and determining the target feature tag from a plurality of feature tags based on the feature tag coefficient includes:
Adding the abnormal quantity and the normal quantity to obtain a total quantity, performing proportion calculation based on the abnormal quantity and the total quantity to obtain an abnormal proportion, and performing calculation based on the abnormal proportion, the preset proportion and the preset quantity to obtain a calculated quantity;
Acquiring the wear weight of the wear tags corresponding to each view tag, and carrying out addition average calculation on the wear weights of the wear tags corresponding to the view tags of the same type to obtain average wear weights;
calculating to obtain characteristic label coefficients of corresponding types based on the average wear weight and the number of abnormal wear labels corresponding to each type of view labels;
And determining the target characteristic label from a plurality of characteristic labels based on the calculated number and the characteristic label coefficient.
Optionally, in a possible implementation manner of the first aspect, the determining, based on the calculated number and the feature tag coefficient, the target feature tag from the plurality of feature tags includes:
Performing descending order sorting treatment on the plurality of feature labels based on the feature label coefficients to obtain corresponding descending order sorting sequences;
sequentially traversing and selecting the feature labels with calculated numbers in a descending order sequencing sequence according to the calculated numbers to serve as the determined target feature labels;
The calculated number, average wear weight and feature tag coefficient are obtained by the following formula,
Wherein xcal is the calculated number, sabn is the abnormal number, snor is the normal number, g is a preset constant value, bpre is a preset ratio, xpre is a preset number,For the i-th type of feature tag coefficient, uj is the loss weight corresponding to the j-th feature tag in the i-th type, H is the upper limit value of the feature quantity extracted from the historical image corresponding to the feature tag in the i-th type, H is the quantity value of the feature quantity extracted from the historical image corresponding to the feature tag in the i-th type, ei is the quantity of abnormal loss tags corresponding to the feature tag in the i-th type, and gqua is the quantity normalization value.
Optionally, in one possible implementation manner of the first aspect, the image sub-model extracts a target view tag and a target wear tag corresponding to the target feature tag, establishes a corresponding image slot in an information extraction template based on the target view tag and the target wear tag, and includes:
extracting a target view label and a target loss label corresponding to a target feature label by an image sub-model, wherein the target loss label is provided with a view loss position corresponding to a target view;
proportional positioning is carried out on the loss position of the target view corresponding to the target view label based on the view loss position;
and establishing a first-level image slot corresponding to the target view at the information extraction template, generating a second-level slot identification area for the target view based on the proportional positioning distribution of all target feature labels in all historical target views, and establishing a corresponding second-level image slot at the first-level image slot.
Optionally, in one possible implementation manner of the first aspect, the proportionally positioning, based on the view loss position, the loss position of the target view corresponding to the target view tag includes:
acquiring a loss view corresponding to a view loss position, carrying out coordinate processing on the target view according to a preset point serving as an origin, and acquiring corner coordinates of a preset corner of the target view;
extracting a first X-axis value and a first Y-axis value corresponding to the angular point coordinates to generate a first X-axis interval value and a first Y-axis interval value corresponding to the angular point coordinates;
acquiring a second X-axis value and a second Y-axis value in the loss view to generate a corresponding second X-axis interval value and a second Y-axis interval value;
And performing proportional positioning of the X axis based on the first X axis interval value and the second X axis interval, and performing proportional positioning of the Y axis based on the first Y axis interval value and the second Y axis interval.
Optionally, in a possible implementation manner of the first aspect, the extracting a corresponding first X-axis value and a corresponding first Y-axis value in the corner coordinate generates a corresponding first X-axis interval value and a corresponding first Y-axis interval value, which includes:
extracting a maximum first X-axis value and a minimum first X-axis value in the angular point coordinates to obtain corresponding first X-axis interval values;
Extracting a maximum first Y-axis value and a minimum first Y-axis value in the angular point coordinates to obtain corresponding first Y-axis interval values;
the obtaining a second X-axis value and a second Y-axis value in the loss view generates a corresponding second X-axis interval value and second Y-axis interval value, including:
extracting a maximum second X-axis value and a minimum second X-axis value in the loss view to obtain a corresponding second X-axis interval value;
Extracting a maximum second Y-axis value and a minimum second Y-axis value in the loss view to obtain a corresponding second Y-axis interval value; the performing the proportional positioning of the X-axis based on the first X-axis interval value and the second X-axis interval, and performing the proportional positioning of the Y-axis based on the first Y-axis interval value and the second Y-axis interval, includes:
Determining a coincident X-axis segment of the second X-axis interval at the first X-axis interval value, and determining a coincident Y-axis segment of the second Y-axis interval at the first Y-axis interval value;
And carrying out proportional positioning according to the loss positions of the coincident X-axis section and the coincident Y-axis section in the target view, and highlighting the proportional positioning area.
In a second aspect of the embodiment of the present invention, there is provided a commodity warehousing interaction system applicable to a multidimensional data model, including:
The system comprises a judging module, a storage module and a data processing module, wherein the judging module is used for enabling a server to send a multi-dimensional data model when judging that the storage demand of a first storage-in end is received, the multi-dimensional data model is provided with a pre-configured storage strategy and an interaction method, and the storage strategy at least comprises at least one of an image extraction sub-strategy, a voice extraction sub-strategy and/or a text extraction sub-strategy;
the analysis module is used for enabling the multidimensional data model to analyze the warehousing requirements, determining a corresponding warehousing strategy and feeding back the corresponding warehousing strategy to a first warehousing end so that the first warehousing end can extract information of a warehousing target according to the warehousing strategy;
The generation module is used for enabling the multidimensional data model to generate a corresponding interaction path based on the interaction method corresponding to the warehousing requirement, wherein an interaction starting point in the interaction path corresponds to the first warehousing end, and other interaction nodes in the interaction path correspond to other warehousing management ends;
the extraction module is used for enabling the multidimensional data model to feed back corresponding warehousing requirements to the server after judging that the interaction data fed back by the warehousing management end and the warehousing data extracted by the first warehousing end are received, so that the server can complete warehousing records.
The technical effects are as follows:
1. The invention improves the warehousing efficiency and accuracy. The multidimensional data model can automatically select the most suitable warehousing strategy according to different characteristics (such as image, voice and text information) of goods. Through a plurality of pre-configured warehousing sub-strategies, the system can quickly make a warehousing plan for different goods and execute the warehouse-in plan, thereby obviously improving warehousing efficiency and accuracy. Meanwhile, the method can generate a corresponding interaction path, and can obviously improve the speed and accuracy of warehouse-in operation and reduce human errors. The flexibility and adaptability of the system are enhanced. According to the method, a multidimensional data model capable of analyzing various data types is introduced, so that the adaptability of the system to different warehouse-in requirements is enhanced. Whether the system is image data, sound or text labels, the system can recognize and adjust the interaction strategy according to specific contents, so that the system can flexibly cope with changeable logistics demands.
2. According to the invention, through the multidimensional data model and the highly customized information extraction template, the automatic process of goods warehousing is greatly optimized, so that the warehousing efficiency and accuracy are remarkably improved. Specifically, the system not only realizes the efficient processing of data in the traditional logistics and warehouse management system, but also improves the information flow and processing consistency through an innovative interaction path generation mechanism. Through the collaborative work of the image sub-model and other components, the system can accurately identify vision and wear information of the warehouse-in articles. For example, the system automatically identifies and selects the most critical views and wear tags by analyzing image features in the historical data, thereby optimizing the accuracy of information extraction. In addition, the model effectively screens out the image features which are most matched with the current warehouse-in object condition through the calculation of the feature label coefficient, and the accuracy and the reliability of warehouse-in data are further improved.
3. According to the customized operation flow, the system can dynamically generate the information extraction templates containing necessary information slots according to different warehousing requirements, and the templates are optimized according to goods characteristics and warehousing strategies. The customized flow not only enhances the flexibility of the system, but also ensures that various goods can be rapidly and accurately processed as required. The information extraction template reduces the operation complexity, so that the first warehouse-in end can efficiently complete tasks even without complex training, and the working efficiency is remarkably improved. In addition, the invention also enhances interactive data flow and tracking capability, and high-efficiency data interaction among a plurality of warehouse-in ends not only optimizes the information flow but also ensures the consistency and consistency of data processing through an innovative interaction path generation mechanism. The design of the system enables the whole warehousing process to be more systematic and traceable, and greatly improves the transparency of operation and the error tracking capability. In addition, the system can feed back interactive data and warehousing requirements to the server in real time, further ensures real-time updating and accurate recording of the data, and has important value for inventory management and future strategy formulation. In summary, the technology remarkably improves the efficiency and accuracy of goods warehousing through deep integration of the multi-dimensional data model and application of the intelligent information extraction template, simultaneously enhances the flexibility and adaptability of the system, and provides an innovative and efficient solution for a modern warehouse management system.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The terms "first," "second," "third," "fourth" and the like in the description and in the claims and in the above drawings, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein.
It should be understood that, in various embodiments of the present invention, the sequence number of each process does not mean that the execution sequence of each process should be determined by its functions and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.
It should be understood that in the present invention, "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements that are expressly listed or inherent to such process, method, article, or apparatus.
It should be understood that in the present invention, "plurality" means two or more. "and/or" is merely an association relationship describing an association object, meaning that there may be three relationships, e.g., a and/or B, may represent: a exists alone, A and B exist together, and B exists alone. The character "/" generally indicates that the context-dependent object is an "or" relationship. "comprising A, B and C", "comprising A, B, C" means that all three of A, B, C are comprised, "comprising A, B or C" means that one of A, B, C is comprised, "comprising A, B and/or C" means that any 1 or any 2 or 3 of A, B, C are comprised.
It should be understood that in the present invention, "B corresponding to a", "a corresponding to B", or "B corresponding to a" means that B is associated with a, from which B can be determined. Determining B from a does not mean determining B from a alone, but may also determine B from a and/or other information. The matching of A and B is that the similarity of A and B is larger than or equal to a preset threshold value.
As used herein, the term "if" may be interpreted as "at … …" or "at … …" or "in response to a determination" or "in response to a detection", depending on the context.
The technical scheme of the invention is described in detail below by specific examples. The following embodiments may be combined with each other, and some embodiments may not be repeated for the same or similar concepts or processes.
Referring to fig. 1, a flow chart of a method for interaction in warehousing of goods applicable to a multi-dimensional data model according to an embodiment of the present invention includes:
s1, when judging that the warehouse-in demand of a first warehouse-in end is received, the server sends the warehouse-in demand to a multidimensional data model, wherein the multidimensional data model is provided with a preset warehouse-in strategy and an interaction method, and the warehouse-in strategy at least comprises at least one of an image extraction sub-strategy, a voice extraction sub-strategy and/or a text extraction sub-strategy.
The first warehouse-in end can be held by warehouse-in personnel, and when the first warehouse-in end has warehouse-in requirements, the first warehouse-in end can be automatically operated by means of the multidimensional data model. The multidimensional data model is a pre-configured data processing model, and can execute corresponding data processing and interaction strategies according to different requirements. The multidimensional data model contains different sub-policies for processing for different types of data (e.g., images, speech, text).
In some embodiments, the server sends the warehousing requirements to a multidimensional data model when judging that the warehousing requirements of the first warehousing end are received, wherein the multidimensional data model has a preconfigured warehousing strategy and an interaction method, and the method comprises the following steps:
S11, when judging that the warehouse-in demand of the first warehouse-in end is received, the server extracts a warehouse-in article label corresponding to the warehouse-in demand and sends the warehouse-in article label to the multi-dimensional data model.
The server first needs to determine whether a warehousing requirement is received from a first warehousing end (such as an interface of an automated logistics system and input of a warehousing personnel), and once the warehousing requirement is confirmed to be received, the server extracts the warehousing article label information related to the requirement. It is worth mentioning that, articles stored in warehouse having one-to-one advance and the warehouse-in article labels are correspondingly arranged.
S12, determining corresponding warehousing strategies and interaction methods based on the warehousing article labels by the multidimensional data model, wherein each warehousing article label has a preset warehousing strategy and interaction method.
It can be understood that different warehouse-in articles have different warehouse-in strategies and interaction methods, and the scheme can be used for judging through the warehouse-in article labels. For example, the article a needs to perform warehouse entry data processing of images and voices; the article B needs to carry out warehouse-in data processing of images and characters; . Different binning strategies are used to process different types of data (e.g., images, speech, text).
Through the warehousing processing flow of the multidimensional data model, the server can efficiently and intelligently process different types of warehousing requirements, and accurate classification and storage of data are ensured. Such a system improves the operating efficiency of the automated warehouse and reduces human error. In addition, the preset strategy and method enable the system to be more elaborate and specialized for the handling of different items.
S2, analyzing the warehousing requirements by the multidimensional data model, determining a customized warehousing strategy, and feeding back the customized warehousing strategy to a corresponding first warehousing end so that the first warehousing end can extract information of a warehousing target according to the warehousing strategy.
The multidimensional data model is an automatic processing model, can analyze the warehouse-in requirements in detail, and generates a customized warehouse-in strategy to guide the first warehouse-in end to perform effective information extraction.
In some embodiments, the multidimensional data model analyzes the warehousing requirements, determines a customized warehousing strategy, and feeds back the customized warehousing strategy to a corresponding first warehousing end, so that the first warehousing end performs information extraction on a warehousing target according to the warehousing strategy, and includes:
S21, extracting a warehousing strategy of the multi-dimensional data model for warehousing the article label, wherein the warehousing strategy at least comprises at least one of an image extraction sub-strategy, a voice extraction sub-strategy and/or a text extraction sub-strategy.
The multidimensional data model first analyzes the warehouse entry item labels and extracts the relevant warehouse entry strategies. It will be appreciated that some articles require the acquisition of image information, such as physical equipment, at the time of warehousing; some articles need to collect voice and/or text when in warehouse entry, and different articles have different preset requirements.
S22, generating a customized information extraction template with information extraction marks based on the type information and the number information of the sub-strategies in the warehousing strategy, and feeding back to the first warehousing end.
The scheme can generate an information extraction template according to the type and quantity information of the sub-strategies, the information extraction template is provided with an information extraction mark, and the template is fed back to the first warehouse-in end to guide the information extraction process.
The generating an information extraction template with information extraction identification based on the category information and the quantity information of the sub-policies in the warehousing strategy, and feeding back the information extraction template to a first warehousing end comprises the following steps:
S221, if the warehousing strategy is judged to comprise an image extraction sub-strategy, the image sub-model in the multidimensional data model acquires the warehousing type corresponding to the warehousing object label, and historical data of the warehousing type is analyzed and processed to obtain historical image extraction characteristics corresponding to the warehousing type.
If the warehousing strategy includes an image extraction sub-strategy, the image sub-model of the multi-dimensional data model will analyze the historical data of the category corresponding to the warehousing item label. Through this analysis, the multi-dimensional data model determines historical image extraction features associated with the binned category. It can be understood that the warehousing requirements of the same warehousing article label are approximately similar, so that the scheme can combine the historical image extraction characteristics to determine the extraction data during the warehousing.
S222, obtaining feature labels corresponding to the extracted features of each historical image, and screening the feature labels to obtain target feature labels, wherein the feature labels at least comprise view labels and wear labels.
And the multidimensional data model acquires feature labels corresponding to the features of each historical image, and filters the feature labels to obtain target feature labels. These feature tags include at least view tags (e.g., top view, side view, etc. of the article) and wear tags (e.g., damage to the article).
The method comprises the steps of obtaining a feature tag corresponding to each historical image extraction feature, screening the feature tag to obtain a target feature tag, wherein the feature tag at least comprises a view tag and a wear tag, and the method comprises the following steps:
Counting the abnormal quantity of abnormal wear-out labels in each view label in all historical image extraction characteristics of the warehouse-in article labels of the same kind, and counting the normal quantity of normal wear-out labels in each view label. The method mainly comprises the steps of counting the abnormal quantity of the abnormal loss labels in each view label, wherein it can be understood that the more the abnormal quantity of the abnormal loss labels is, the greater the probability of easily causing problems of the image corresponding to the corresponding view label is, and the easily-lost surface of the equipment can be found; meanwhile, the scheme also counts the normal number of the loss labels in each view label. It can be understood that the more the normal number of the loss labels is normal, the smaller the probability that the content corresponding to the image corresponding to the corresponding view label is easy to cause a problem, and the less likely the device is lost can be found.
And generating a wear weight of each wear tag based on the description information of the wear tag, wherein each description information has a preset wear weight. The description information of the wear-out tag may be, for example, slight wear, medium wear, severe wear, etc. Each wear-out tag has a preset wear-out weight according to the description information, which can be preset by warehouse personnel in combination with experience, wherein the wear-out weight corresponding to serious wear-out can be larger than the wear-out weight corresponding to medium wear-out, and the wear-out weight corresponding to medium wear-out can be larger than the wear-out weight corresponding to slight wear-out.
And calculating based on the abnormal quantity, the normal quantity and the wear weight to obtain feature tag coefficients of the extracted features of the corresponding historical images, and determining a target feature tag from a plurality of feature tags based on the feature tag coefficients. The scheme calculates feature tag coefficients based on the calculation results of the abnormal number, the normal number and the wear weight, and finally determines the target feature tag by using the feature tag coefficients. Through the mode, the first warehouse-in end can efficiently and automatically determine the warehouse-in target.
The calculating based on the abnormal number, the normal number and the wear weight to obtain a feature tag coefficient of the extracted feature of the corresponding historical image, and determining a target feature tag from a plurality of feature tags based on the feature tag coefficient comprises:
Adding the abnormal quantity and the normal quantity to obtain a total quantity, calculating the proportion based on the abnormal quantity and the total quantity to obtain an abnormal proportion, and calculating the proportion based on the abnormal proportion, the preset proportion and the preset quantity to obtain a calculated quantity. It will be appreciated that the greater the number of anomalies, the greater the corresponding anomaly ratio. There will be logic in the subsequent formulas that correlates to the calculated number. The number of computations refers to the number of corresponding views, for example, 3 views in 6 views, and thus the larger the anomaly ratio is, the larger the corresponding number of computations is required.
And acquiring the wear weights of the wear tags corresponding to each view tag, and carrying out addition average calculation on the wear weights of the wear tags corresponding to the view tags of the same type to obtain average wear weights. It will be appreciated that the present solution calculates the average wear weight by averaging the wear weights of each view, for example, the left view corresponds to an average wear weight and the right view corresponds to an average wear weight.
And calculating the characteristic label coefficient of the corresponding type based on the average wear weight and the number of the abnormal wear labels corresponding to each type of view labels. It should be noted that, the greater the number of abnormal wear-out labels corresponding to the view labels of the corresponding types, the greater the average wear-out weight, the greater the characteristic label coefficient corresponding to the corresponding types, and the greater the characteristic label coefficient, the higher the degree of abnormality is.
And determining the target characteristic label from a plurality of characteristic labels based on the calculated number and the characteristic label coefficient. In order to comprehensively calculate the number and the characteristic label coefficient, the target characteristic label is obtained.
Wherein the determining the target feature tag from the plurality of feature tags based on the calculated number, feature tag coefficients, comprises:
And performing descending order sorting processing on the plurality of feature labels based on the feature label coefficients to obtain corresponding descending order sorting sequences. It will be appreciated that in the descending order of the above, the higher the feature tag coefficient, and the higher the degree of anomaly.
And sequentially traversing and selecting the feature labels with calculated numbers in a descending order sequencing sequence according to the calculated numbers to serve as the determined target feature labels. After determining the calculated number, the scheme needs to find the corresponding number of feature tags in the descending order sorting sequence in combination with the calculated number, and the feature tags serve as the determined target feature tags, so that the views needing to be acquired are determined.
The calculated number, average wear weight and feature tag coefficient are obtained by the following formula,
Wherein xcal is the calculated number, Sabn is the abnormal number, the normal number Snor is the normal number, g is a preset constant value, bpre is a preset ratio, xpre is a preset number,For the i-th type of feature tag coefficient, uj is the loss weight corresponding to the j-th feature tag in the i-th type, H is the upper limit value of the feature quantity extracted from the historical image corresponding to the feature tag in the i-th type, H is the quantity value of the feature quantity extracted from the historical image corresponding to the feature tag in the i-th type, ei is the quantity of abnormal loss tags corresponding to the feature tag in the i-th type, and gqua is the quantity normalization value.
In the above formula for obtaining the calculated number,Representing the abnormal proportion, wherein the larger the abnormal quantity is, the higher the abnormal proportion is, finally, combining the difference value with the preset proportion to obtain a difference coefficient, and then, performing offset adjustment on the preset quantity to obtain a corresponding calculated quantity, wherein the preset quantity can be 3 or can be preset by a worker. If it isThe abnormal proportion is smaller than the preset proportion, and the preset quantity can be used as the calculated quantity at the moment; if it isThe abnormal proportion is larger than or equal to the preset proportion, at this time, the preset quantity can be increased and adjusted to obtain the calculated quantity, and if the calculated quantity is not an integer, the upward rounding process can be performed.
In the above formula for deriving the characteristic tag coefficients,Representing average wear-out weight, the larger the number ei of abnormal wear-out labels corresponding to the feature labels in the ith type is, the larger the degree of abnormality of the corresponding view type is, and the larger the finally obtained feature label coefficient is; the number normalization value may be preset by a worker.
S223, extracting a target view label and a target loss label corresponding to the target feature label by the image sub-model, and establishing a corresponding image slot position on the information extraction template based on the target view label and the target loss label. It should be noted that, the target feature labels may include target view labels and target wear labels, the target view labels may correspond to respective views, such as front view, side view, top view, etc., each target view label may have a corresponding target wear label, and the target wear labels may be views of corresponding wear in the front view, side view, top view, etc.
The target view label is, for example, a front view of the device a, the target wear label is, for example, a partial part in the front view, which is relatively easy to wear, and then the corresponding image slot is established in the information extraction template by combining the data. Through the process, the first warehouse-in end can efficiently extract accurate information of the warehouse-in target. The information extraction template provides a mechanism for the first warehousing end to accurately extract and classify relevant information of the warehoused article, such as image features and wear conditions. This helps optimize inventory management, improves the overall efficiency of the automated warehouse system, and reduces the likelihood of errors.
The method for extracting the target view label and the target wear label corresponding to the target feature label by the image sub-model comprises the steps of:
And the image sub-model extracts a target view label and a target loss label corresponding to the target feature label, wherein the target loss label is provided with a view loss position corresponding to the target view. Wherein the view loss position is for example the position of a certain partial section in the front view.
And proportionally positioning the loss position of the target view corresponding to the target view label based on the view loss position. This solution requires locating the data to be acquired in conjunction with view loss locations.
The proportional positioning of the loss position of the target view corresponding to the target view tag based on the view loss position includes:
Obtaining a loss view corresponding to the view loss position, carrying out coordinate processing on the target view according to the preset point serving as an origin, and obtaining the corner coordinates of the preset corner of the target view. Taking the loss view as a front view example, a preset point is, for example, the lower left corner of the front view, carrying out coordinate processing on the front view, identifying four preset corner points of the equipment outline in the front view, and then determining the corner point coordinates of the preset corner points.
In some embodiments, if the corresponding target device is a regular rectangle, the corresponding preset corner points are four corner points of the rectangle; in other embodiments, if the target device corresponding to the front view is determined to be an irregular rectangle, the coordinates of the four preset corner points do not correspond to each other, so that the extremum coordinates corresponding to the target device in the target view need to be obtained at this time, and the corresponding corner point coordinates are generated according to the maximum X value, the minimum X value, the maximum Y value, and the minimum Y value in the extremum coordinates, where the target view is a proposed target view, and includes the corresponding target device but does not completely correspond to the target device.
Referring to fig. 2, corresponding first X-axis values and first Y-axis values in the corner coordinates are extracted to generate corresponding first X-axis interval values and first Y-axis interval values. After the angular point coordinates are determined, the scheme needs to find a first X-axis value and a first Y-axis value corresponding to the angular point coordinates, wherein the first X-axis value refers to a data value in the transverse direction corresponding to the equipment, and the first Y-axis value refers to a data value in the longitudinal direction corresponding to the equipment.
The extracting the corresponding first X-axis value and first Y-axis value in the angular point coordinates to generate a corresponding first X-axis interval value and first Y-axis interval value includes:
And extracting the largest first X-axis value and the smallest first X-axis value in the angular point coordinates to obtain corresponding first X-axis interval values. The largest first X-axis value and the smallest first X-axis value may form a first X-axis interval value.
And extracting the largest first Y-axis value and the smallest first Y-axis value in the angular point coordinates to obtain corresponding first Y-axis interval values. The largest first Y-axis value and the smallest first Y-axis value may form a first X-axis interval value.
And obtaining a second X-axis value and a second Y-axis value in the loss view to generate a corresponding second X-axis interval value and a second Y-axis interval value.
And performing proportional positioning of the X axis based on the first X axis interval value and the second X axis interval value, and performing proportional positioning of the Y axis based on the first Y axis interval value and the second Y axis interval value.
And after the first Y-axis interval value and the second Y-axis interval value are obtained, the first Y-axis interval value and the second Y-axis interval value are combined to perform Y-axis proportional positioning.
In some embodiments, the acquiring the second X-axis value and the second Y-axis value in the loss view generates a corresponding second X-axis interval value and second Y-axis interval value, comprising:
And extracting the largest second X-axis value and the smallest second X-axis value in the loss view to obtain corresponding second X-axis interval values. The largest second X-axis value and the smallest second X-axis value may form a second X-axis interval value. The first coordinate corresponds to the target view, the loss view is a local area in the target view, the loss view is a view determined by personnel, the view is a standard rectangular view, and the simulated rectangle processing is not needed. In a practical application scenario, the wear-out view may include a plurality of parts, and generally, the wear-out part is located in the middle of the wear-out view, and the wear-out view may further include other parts that are not worn out.
And extracting the largest second Y-axis value and the smallest second Y-axis value in the loss view to obtain corresponding second Y-axis interval values. The largest second Y-axis value and the smallest second Y-axis value may form a second Y-axis interval value.
The performing the proportional positioning of the X-axis based on the first X-axis interval value and the second X-axis interval value, and performing the proportional positioning of the Y-axis based on the first Y-axis interval value and the second Y-axis interval value, includes:
And determining a coincident X-axis segment of the second X-axis interval value at the first X-axis interval value, and determining a coincident Y-axis segment of the second Y-axis interval value at the first Y-axis interval value. The scheme needs to find out repeated data of the second X-axis interval value and the first X-axis interval value and meanwhile find out coincident data of the Y-axis.
And carrying out proportional positioning according to the loss positions of the coincident X-axis section and the coincident Y-axis section in the target view, and highlighting the proportional positioning area. Because different staff have different shooting habits, when shooting a certain device, the proportion of the devices shot by different staff in the image is different, but the proportion relation between each part in the device and the device is certain, so the invention needs to adopt the proportion type positioning to determine the corresponding part in one device, and the proportion type can enable the rapid positioning of the parts according to the proportion relation between the parts and the device even though the shot images have different sizes.
And establishing a first-level image slot corresponding to the target view at the information extraction template, generating a second-level slot identification area for the target view based on the proportional positioning distribution of all target feature labels in all historical target views, and establishing a corresponding second-level image slot at the first-level image slot.
Referring to fig. 3, the first-level image slot refers to a slot corresponding to an equipment image in the view, for example, a slot corresponding to a front view image, a slot corresponding to a top view image, and the like; the above-mentioned two-level image slot refers to a slot of a certain local area in the view.
S23, the first warehouse-in end interacts with the information extraction template based on the information extraction mark, and information extraction is carried out on the warehouse-in target.
The first warehouse-in end of the scheme can combine the information extraction mark and the information extraction template to extract required data of warehouse-in articles.
On the basis of the foregoing embodiment, the generating, based on the type information and the number information of the sub-policies in the warehousing policy, an information extraction template with an information extraction identifier, and feeding back the information extraction template to a first warehousing terminal includes:
If the warehousing strategy is judged to comprise a voice extraction sub-strategy and/or a text extraction sub-strategy, a first type slot corresponding to the voice extraction sub-strategy and/or the text extraction sub-strategy is built in the information extraction template, and the first type slot is a slot which is necessary to fill. It can be appreciated that when there is a voice extraction requirement and/or a text extraction requirement, the embodiment needs to generate a corresponding first type slot on the information extraction template, including a first type slot corresponding to voice and/or a first type slot corresponding to text. The corresponding type of data may then be filled in the first type of slots. The voice-related data may be, for example, transportation data input at the first warehouse-in end, and the text-related data may be, for example, time data input at the first warehouse-in end.
If the secondary slot position recognition area exists, establishing a voice extraction sub-strategy and/or a text extraction sub-strategy and a second type slot position corresponding to the secondary slot position recognition area, wherein the second type slot position is a filling slot position. It can be understood that the scheme also establishes a corresponding second type slot in the second-level slot identification area so as to allow warehousing personnel to add remark data to the content corresponding to the second-level slot area. The meaning that the second type of slot is a filling slot means that the warehouse-in personnel can choose not to add corresponding data.
And S3, generating a corresponding interaction path by the multidimensional data model based on the interaction method corresponding to the warehousing requirement, wherein an interaction starting point in the interaction path corresponds to the first warehousing end, and other interaction nodes in the interaction path correspond to other warehousing management ends. It can be understood that the start point of the interaction path corresponds to the first entering end, while other interaction nodes on the path correspond to other entering management ends, i.e. a plurality of entering ends may be corresponding to one interaction path.
In some embodiments, the multidimensional data model generates a corresponding interaction path based on an interaction method corresponding to the warehousing requirement, an interaction starting point in the interaction path corresponds to the first warehousing end, and other interaction nodes in the interaction path correspond to other warehousing management ends, including:
s31, the multidimensional data model generates corresponding interaction paths based on interaction methods corresponding to the warehousing requirements, and each warehousing requirement has a preset interaction path.
It should be noted that the value of the equipment is different for different equipments, and the personnel required for warehousing are different during warehousing, for example, the equipment A requires personnel 1-3-4 for warehousing, and the equipment B requires personnel 2-3-5 for warehousing, so that different interaction paths exist.
S32, the information extraction template is sent to other warehouse-in management terminals, the other warehouse-in management terminals add corresponding voice extraction sub-strategies and/or text extraction sub-strategies to the information extraction template, and the voice extraction sub-strategies and/or text extraction sub-strategies are correspondingly arranged with the generated proportional positioning areas.
If other warehouse-in management terminals need to add voice or text information, corresponding voice extraction sub-strategies and/or text extraction sub-strategies can be added to the information extraction template, then the voice extraction sub-strategies and/or text extraction sub-strategies are correspondingly arranged with corresponding areas, and the needed remark information can be added later.
And S4, after judging that the multi-dimensional data model receives the warehouse-in data extracted by the first warehouse-in end and the interactive data fed back by the warehouse-in management end, feeding back corresponding warehouse-in requirements to the server so as to enable the server to finish warehouse-in records.
The multidimensional data model is used as a key information processing and judging center and is connected with the first warehouse-in end, the warehouse-in management end and the server, and a seamless data link is created to ensure the accuracy and efficiency of the warehouse-in process. In this way, the multidimensional data model is effectively applied as an information integration and decision support tool in intelligent logistics and warehouse management systems. The automatic level of the warehousing process is improved, human errors are reduced, the whole warehousing process is quickened, and real-time and accurate inventory management is provided.
Referring to fig. 4, a schematic structural diagram of a commodity warehousing interaction system suitable for a multi-dimensional data model according to an embodiment of the present invention includes:
The system comprises a judging module, a storage module and a data processing module, wherein the judging module is used for enabling a server to send a multi-dimensional data model when judging that the storage demand of a first storage-in end is received, the multi-dimensional data model is provided with a pre-configured storage strategy and an interaction method, and the storage strategy at least comprises at least one of an image extraction sub-strategy, a voice extraction sub-strategy and/or a text extraction sub-strategy;
the analysis module is used for enabling the multidimensional data model to analyze the warehousing requirements, determining a corresponding warehousing strategy and feeding back the corresponding warehousing strategy to a first warehousing end so that the first warehousing end can extract information of a warehousing target according to the warehousing strategy;
The generation module is used for enabling the multidimensional data model to generate a corresponding interaction path based on the interaction method corresponding to the warehousing requirement, wherein an interaction starting point in the interaction path corresponds to the first warehousing end, and other interaction nodes in the interaction path correspond to other warehousing management ends;
the extraction module is used for enabling the multidimensional data model to feed back corresponding warehousing requirements to the server after judging that the interaction data fed back by the warehousing management end and the warehousing data extracted by the first warehousing end are received, so that the server can complete warehousing records.
The present invention also provides a storage medium having stored therein a computer program for implementing the methods provided by the various embodiments described above when executed by a processor.
The storage medium may be a computer storage medium or a communication medium. Communication media includes any medium that facilitates transfer of a computer program from one place to another. Computer storage media can be any available media that can be accessed by a general purpose or special purpose computer. For example, a storage medium is coupled to the processor such the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium may be integral to the processor. The processor and the storage medium may reside in an Application SPECIFIC INTEGRATED Circuits (ASIC). In addition, the ASIC may reside in a user device. The processor and the storage medium may reside as discrete components in a communication device. The storage medium may be read-only memory (ROM), random-access memory (RAM), CD-ROMs, magnetic tape, floppy disk, optical data storage device, etc.
The present invention also provides a program product comprising execution instructions stored in a storage medium. The at least one processor of the device may read the execution instructions from the storage medium, the execution instructions being executed by the at least one processor to cause the device to implement the methods provided by the various embodiments described above.
In the above embodiments of the terminal or the server, it should be understood that the Processor may be a central processing unit (english: central Processing Unit, abbreviated as CPU), or may be other general purpose processors, digital signal processors (english: DIGITAL SIGNAL Processor, abbreviated as DSP), application specific integrated circuits (english: application SPECIFIC INTEGRATED Circuit, abbreviated as ASIC), or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of a method disclosed in connection with the present invention may be embodied directly in a hardware processor for execution, or in a combination of hardware and software modules in a processor for execution.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.