Discrimination standard generation method, surface quality inspection method, storage medium and program product

Abstract

The present application provides a method for generating a discrimination standard, a surface quality inspection method, a storage medium and a program product. The method for generating a surface quality discrimination standard includes: obtaining surface defects contained in the surfaces of multiple sample objects, and sample labels corresponding to each sample object, wherein the sample labels are used to characterize whether the sample objects meet preset standards; based on a predetermined surface defect category, according to the surface defects contained in the surfaces of the sample objects, obtaining defect feature information of each of the sample objects in the surface defect category; establishing a sample data set corresponding to the surface defect category according to the defect feature information of each of the sample objects in the surface defect category and the sample labels of each of the sample objects; performing rule learning based on the sample data set corresponding to the surface defect category to obtain a surface quality discrimination standard corresponding to the surface defect category.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of computer technology, and in particular to a method for generating a discrimination standard, a surface quality inspection method, a storage medium, and a program product. Background Art

[0002] In the production and processing of industrial products, due to the process or equipment, there will be certain defects on the surface of industrial products. For example, the surface of the steel coil may have inclusions, scars, peeling, iron oxide scale, warping, scratches, etc. Therefore, it is necessary to conduct quality inspection on the surface of industrial products.

[0003] Generally, the surface of industrial products is rated according to certain industry standards. For example, based on the preset industry standards, the surface quality of steel coils is rated according to the quantitative description of defects such as the number, length, width, area, and diameter of different types of defects on the surface of steel coils. However, these industry standards are generally customized by the manufacturers of industrial products based on experience, which are relatively lagging and difficult to modify. Summary of the invention

[0004] In view of this, an embodiment of the present application provides a solution for generating a surface quality judgment standard to at least partially solve the above-mentioned problem.

[0005] According to a first aspect of an embodiment of the present application, a method for generating a surface quality judgment standard is provided, comprising: obtaining surface defects contained in the surfaces of multiple sample objects, and sample labels corresponding to each of the sample objects, the sample labels being used to characterize whether the sample objects meet preset standards; based on a predetermined surface defect category, obtaining defect feature information of each of the sample objects in the surface defect category according to the surface defects contained in the surfaces of the sample objects; establishing a sample data set corresponding to the surface defect category according to the defect feature information of each of the sample objects in the surface defect category and the sample labels of each of the sample objects; performing rule learning based on the sample data set corresponding to the surface defect category to obtain a surface quality judgment standard corresponding to the surface defect category.

[0006] According to a second aspect of an embodiment of the present application, a surface quality inspection method for a target object is provided, comprising: obtaining defect feature information of the target object in each surface defect category; based on a surface quality judgment standard corresponding to each surface defect category, judging whether the target object meets a preset standard according to the defect feature information of the target object in each surface defect category, the surface quality judgment standard is determined according to the method described above.

[0007] According to a third aspect of an embodiment of the present application, a steel coil surface quality inspection method is provided, comprising: obtaining defect feature information of a target steel coil in each surface defect category; based on a surface quality judgment standard corresponding to each surface defect category, judging whether the target steel coil meets a preset standard according to the defect feature information of the target steel coil in each surface defect category, the surface quality judgment standard being determined according to the method described above.

[0008] According to a fourth aspect of an embodiment of the present application, a computer storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the method described above is implemented.

[0009] According to a fifth aspect of an embodiment of the present application, a computer program product is provided, comprising computer instructions, wherein the computer instructions instruct a computing device to perform operations corresponding to the method described above.

[0010] According to the scheme provided by the embodiment of the present application, surface defects contained in the surfaces of multiple sample objects and sample labels corresponding to the sample objects can be obtained, and the sample labels are used to characterize whether the sample objects meet the preset standards; based on the predetermined surface defect categories, defect feature information of each sample object in the surface defect category is obtained according to the surface defects contained in the surface of the sample objects; and a sample data set corresponding to the surface defect category is established according to the defect feature information of each sample object in the surface defect category and the sample labels of each sample object; rule learning is performed according to the sample data set corresponding to the surface defect category to obtain the surface quality judgment standard corresponding to the surface defect category. By establishing a sample data set corresponding to the preset surface defect category according to the surface defects and sample labels of the sample objects, rule learning can be performed based on this, so that according to the surface defects actually included in the sample objects and whether they meet the preset standards, the surface quality judgment standard for judging whether the target object meets the preset standard is learned, and the surface quality judgment standard is automatically and quantifiablely generated. Compared with the judgment standard manually formulated in the past, the accuracy and recall rate of surface quality inspection can be improved, and the surface quality judgment standard suitable for different scenarios can be produced by changing the sample labels, which is convenient for meeting the needs of different users. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the embodiments of the present application. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0012] Figure 1 A schematic diagram of a flow chart of a method for generating a surface quality judgment standard provided in an embodiment of the present application;

[0013] Figure 2A A schematic diagram of a flow chart of a method for generating a surface quality judgment standard provided in an embodiment of the present application;

[0014] Figure 2B A schematic diagram of a decision tree generated according to an embodiment of the present application;

[0015] Figure 2C A schematic diagram of the principle of an original criterion for judging the quality of a filtering surface provided in an embodiment of the present application;

[0016] Figure 2D A schematic diagram of the principle of semantic filtering of the original surface quality judgment standard provided in an embodiment of the present application;

[0017] Figure 3 A schematic diagram of a process for surface quality inspection of a target object provided in an embodiment of the present application;

[0018] Figure 4 A schematic diagram of a process for inspecting the surface quality of a steel coil provided in an embodiment of the present application;

[0019] Figure 5 It is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0020] In order to enable those skilled in the art to better understand the technical solutions in the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments in the embodiments of the present application should fall within the scope of protection of the embodiments of the present application.

[0021] In order to explain the solution provided by this embodiment in more detail, the usage scenario in this application is first exemplified below.

[0022] Take steel coils as an example. Due to process and equipment reasons, steel coils will have certain surface defects during the production process. Common surface defects of steel coils can be classified into dozens of types according to their morphology and causes, including inclusions, scars, peeling, iron oxide scale, warping, scratches, scrapes, etc.

[0023] Different users or usage scenarios have different requirements for the surface of steel coils, which requires steel coil manufacturers to accurately inspect the surface of steel coils and classify the steel coils into different levels or categories based on the inspection results. Generally, the industry standards for surface inspection of steel coils are generally determined by negotiation among multiple steel coil manufacturers to ensure that the standards within the industry are relatively uniform. However, it is precisely because of this that the standards for steel coil quality inspection are difficult to update and cannot meet the diverse needs of users.

[0024] For example, it is possible that some steel coils that meet certain usage requirements are reworked or sealed because they do not meet the rigid grade determination rules, resulting in production waste; and some steel coils that do not meet usage requirements are mistakenly judged as qualified and enter the market because of inappropriate surface quality judgment standards, causing customer complaints and economic and reputational losses to the company.

[0025] To this end, an embodiment of the present application provides a method for generating a surface quality judgment standard, which can quickly and accurately generate a surface quality judgment standard.

[0026] Figure 1 A flow chart of a method for generating a surface quality judgment standard provided in an embodiment of the present application is shown in the figure, which includes:

[0027] S101, obtaining surface defects contained in the surfaces of a plurality of sample objects, and sample labels corresponding to the sample objects.

[0028] In this embodiment, the sample object may be an industrial product with a determined sample label, such as a steel coil, aluminum material, etc., which is not limited in this embodiment.

[0029] The surface defects contained in the sample object may include, for example, inclusions, scars, peeling, iron oxide scale, peeling, scratches, etc., and may also include the location, size, depth, etc. of each surface defect. The surface defects can be specifically detected by a surface inspection instrument.

[0030] The sample label is used to characterize whether the sample object meets the preset standard. For example, the preset standard may be, for example, whether it meets user requirements, whether it meets a certain industry standard, etc. This embodiment does not limit this, and technicians in this field can determine it according to needs.

[0031] S102: Based on a predetermined surface defect category, and according to the surface defects contained on the surface of the sample object, obtain defect feature information of each of the sample objects in the surface defect category.

[0032] The preset defect categories may specifically be inclusions, scars, peeling, iron oxide scale, warping, scratches, scrapes, etc. shown in the above steps, or may be surface defect categories determined by technicians in this field, which are not limited in this embodiment.

[0033] Each surface defect category may correspond to a defect feature, such as the number of defects, defect size, etc. The defect feature information of the sample object in the surface defect category is the feature value of each defect feature of the surface defect category corresponding to the sample object. For example, the defect feature information of a sample object in the surface defect category may include: the number of defects is 10, the maximum length of the defect is 2 mm, etc.

[0034] S103 . Establish a sample data set corresponding to the surface defect category according to the defect feature information of each sample object in the surface defect category and the sample label of each sample object.

[0035] In this embodiment, the sample data set may include: a plurality of sample objects having the surface defect category, defect feature information of each sample object in the surface defect category, and a sample label corresponding to each sample object.

[0036] In this embodiment, when there are multiple surface defect categories, the number of sample data sets may also be multiple, corresponding one-to-one to the surface defect categories.

[0037] S104 , performing rule learning according to the sample data set corresponding to the surface defect category to obtain the surface quality judgment standard corresponding to the surface defect category.

[0038] For specific methods of rule learning, please refer to relevant technologies and will not be described in detail here.

[0039] The solution provided in this embodiment can be based on obtaining surface defects contained in the surfaces of multiple sample objects and sample labels corresponding to each sample object, wherein the sample labels are used to characterize whether the sample objects meet the preset standards; based on the predetermined surface defect categories, defect feature information of each sample object in the surface defect category is obtained according to the surface defects contained in the surface of the sample object; and a sample data set corresponding to the surface defect category is established according to the defect feature information of each sample object in the surface defect category and the sample labels of each sample object; rule learning is performed according to the sample data set corresponding to the surface defect category to obtain the surface quality judgment standard corresponding to the surface defect category. By using the surface defects and sample labels of the sample objects, a sample data set corresponding to the preset surface defect category can be established, and rule learning can be performed based on this, so that according to the surface defects actually included in the sample objects and whether they meet the preset standards, the surface quality judgment standard for judging whether the target object meets the preset standards is learned, and the surface quality judgment standard is automatically and quantifiablely generated. Compared with the judgment standard manually formulated in the past, the accuracy and recall rate of surface quality inspection can be improved, and the surface quality judgment standard suitable for different scenarios can be produced by changing the sample labels, which is convenient for meeting the needs of different users.

[0040] The method for generating the surface quality judgment standard of this embodiment can be executed by any appropriate electronic device with data processing capability, including but not limited to: a server, a mobile terminal (such as a mobile phone, a PAD, etc.) and a PC, etc.

[0041] Figure 2A A flow chart of a method for generating a surface quality judgment standard provided in an embodiment of the present application is shown in the figure, which includes:

[0042] S201, obtaining surface defects contained in surfaces of a plurality of sample objects and sample labels corresponding to the sample objects.

[0043] The sample label is used to indicate whether the sample object meets a preset standard.

[0044] Optionally, in this embodiment, since user needs are more variable, and the rating of the target object is also to provide users with objects of corresponding levels to meet user needs, in this embodiment, the sample label is specifically used to characterize whether the sample object meets a preset standard, and the preset standard is the product surface quality standard required by the user who purchases the sample object. Similarly, the preset standard can also be a preset surface level standard.

[0045] When the preset standard is the product surface quality standard required by the user who purchases the sample object, if the user continues to purchase a certain object, such as a steel coil, etc., the sample label of the newly added sample object is determined based on the sample objects added within the preset time period and the feedback information from the user who purchased the sample object; an updated sample data set corresponding to the surface defect category is established based on the defect feature information of the sample objects added within the preset time period on the surface defect category and the sample label; rule learning is performed based on the updated sample data set corresponding to the surface defect category to update the surface quality judgment standard corresponding to the surface defect category according to the learning result.

[0046] Thus, the surface quality judgment standard can be updated according to the user transaction evaluation obtained during the transaction with the user within the preset time period, so that the surface quality judgment standard is updated in time with the user's needs, so that the surface quality judgment standard can be updated according to the updated surface quality judgment standard.

[0047] S202 , clustering the surface defects according to the morphological features of the surface defects respectively included in the surfaces of the plurality of sample objects.

[0048] In this embodiment, the morphological features are visual features of surface defects, which may specifically include information such as the shape and depth of each surface defect, which is not limited in this embodiment. Surface defects with the same or similar morphology may be classified into one category according to the morphological features.

[0049] S203: Determine the surface defect category according to the clustering result.

[0050] In this embodiment, the clustering result may include multiple clusters, each cluster may correspond to a surface defect feature, and in this step, a data list of surface defect categories may be obtained.

[0051] S204: Based on the determined surface defect category, and according to the surface defects contained on the surfaces of the sample objects, obtain defect feature information of each of the sample objects in the surface defect category.

[0052] Optionally, in this embodiment, the sample object is a steel coil or an aluminum material, the surface defect information is a characteristic value corresponding to each of the defect characteristics, and the defect characteristics corresponding to the surface defect category include at least one of the following: number of defects, number of defects per meter, number of defects excluding the head and tail, cumulative length, cumulative length ratio, cumulative width, cumulative area ratio, maximum defect length, maximum defect width, maximum defect area, maximum distance from the edge, maximum distance from the head, and maximum distance from the tail.

[0053] Specifically, referring to the following Table 1, 13 defect features included in the surface defect features are shown. Since the steel coil or aluminum material has upper and lower surfaces, the final number of defect features may be 13*2=26.

[0054] As shown in Table 1, the defect characteristics of the surface defect category can be specifically divided into two categories, namely, cumulative defect characteristics and extreme defect characteristics. The cumulative defect characteristics are the cumulative values ​​of the surface defect characteristics of this type that exist in a certain sample object, and the extreme defect characteristics are the limit values ​​of the surface defect characteristics of this type that exist in a certain sample object. The characteristic code of the defect characteristic can specifically be the code corresponding to the defect characteristic used when generating the surface defect judgment standard.

[0055]

[0056] Table 1

[0057] Specifically, if there are n surface defect categories, each sample object may correspond to the surface defect category in Table 1 above one by one, and the defect feature information of the sample object may specifically be the feature value corresponding to each defect feature in Table 1 above.

[0058] S205 . Establish a sample data set corresponding to the surface defect category according to the defect feature information of each of the sample objects in the surface defect category and the sample label of each of the sample objects.

[0059] S206: Divide the sample data set corresponding to the surface defect category into a training set and a validation set.

[0060] The specific method of dividing the training set and the validation set can be referred to in the relevant technology and will not be described in detail here.

[0061] S207, performing rule learning according to the training set based on a random forest algorithm to obtain an original surface quality discrimination standard corresponding to the surface defect category;

[0062] Random forest: In machine learning, a random forest is a classifier that contains multiple decision trees, and the category of its output is determined by the mode of the categories output by individual trees. In this embodiment, the output of the random forest is not a category, but the original discrimination criteria of surface quality.

[0063] Specifically, in this embodiment, defect features can be randomly extracted from various defect features corresponding to the surface defect category; sampling with replacement is performed on the training set, and the feature threshold corresponding to the defect feature is determined according to the sampling result; node splitting is performed according to the defect feature and the feature threshold to generate a decision tree; a random forest is established based on the generated decision tree, and the original surface quality discrimination standard corresponding to the surface defect category is determined according to the defect features and feature thresholds corresponding to the leaf nodes in the random forest.

[0064] Decision tree: Based on the known probability of occurrence of various situations, a decision tree is constructed to find the probability that the expected value of the net present value is greater than or equal to zero. This is a graphical method of intuitively applying probability analysis. Since this decision branch is drawn as a graph that looks like a tree branch, it is called a decision tree. In machine learning, a decision tree is a prediction model that represents a mapping relationship between object attributes and object values.

[0065] In this embodiment, multiple decision trees can be generated to form a random forest. The decision tree is composed of multiple nodes. The lowest layer is called the leaf node of the decision tree. When generating a decision tree, for a node in the decision tree, the defect feature can be randomly extracted from the defect features corresponding to the surface defect category; the training set is sampled with replacement, and the feature threshold corresponding to the defect feature is determined according to the sampling result; the node is split according to the defect feature and the feature threshold. If the node is classified, two child nodes corresponding to the node can be generated. If the node is not split, it means that the node is a leaf node of the decision tree.

[0066] For example, see Figure 2B , shows a schematic diagram of a generated decision tree, where the defect features are below each node and the range determined based on the feature threshold for splitting the node is above the node. The sample label statistics of the randomly sampled training set are shown in the node.

[0067] by Figure 2B Taking the top node in as an example, the corresponding randomly extracted defect feature is "defect quantity". By sampling the training set with replacement, 12 target objects can be obtained, of which 10 target objects correspond to the sample label of "customer acceptance" and 2 correspond to the sample label of "customer rejection". According to the sampling results, the feature threshold corresponding to the defect feature is determined to be 1000, and the node can be split according to the defect quantity ≤ 1000 and the defect quantity > 1000 to obtain two child nodes.

[0068] against Figure 2BThe child node corresponding to "the number of defects ≤ 1000" shown on the left side can randomly extract the defect feature "defect cumulative length" again, and perform sampling with replacement on the training set to obtain 9 target objects, of which 8 target objects correspond to the sample label of "customer acceptance" and 1 corresponds to the sample label of "customer non-acceptance". According to the sampling results, the feature threshold corresponding to the defect feature is determined to be 10000, and the node can be split according to the defect cumulative length ≤ 10000 and the defect cumulative length > 10000 to obtain two child nodes. The corresponding criteria of these two child nodes are "the number of defects ≤ 1000 and the defect cumulative length ≤ 10000" and "the number of defects ≤ 1000 and the defect cumulative length > 10000" respectively.

[0069] When a node stops splitting, the node can be determined as a leaf node of the decision tree, and its corresponding judgment standard is determined as the original judgment standard of surface quality, for example Figure 2B Rule1 and rule2 are shown.

[0070] Optionally, in this embodiment, since multiple original surface quality discrimination criteria can be generated by random forest, in this implementation, when executing subsequent steps, the original surface quality discrimination criteria can be first screened, filtered and deduplicated.

[0071] See also Figure 2C , shows a schematic diagram of the principle of filtering the original discrimination standard of surface quality. As shown in the figure, when filtering, OOB can be used to generate the accuracy Precision and recall Recall to filter the original discrimination standard of surface quality. OOB refers to out-of-bag samples. Each original discrimination standard of surface quality will have a minimum threshold for the Precision / Recall predicted by the test set. Only those original discrimination standards of surface quality that have reached the threshold in OOB will be retained in this step. The two thresholds are in an "and" relationship, that is, Precision / Recall need to be satisfied at the same time. OOB: out of bag, represents out-of-bag samples, that is, samples that were not sampled by random forest during sampling, and are often used to verify model accuracy.

[0072] like Figure 2C As shown in the figure, the "data not involved in training" on the left is the test set, rule1 and rule2 are the original surface quality judgment standards, and OOB can generate prediction indicators corresponding to each original surface quality judgment standard, namely, accuracy Precision and recall Recall, and can filter out the original surface quality judgment standards whose accuracy Precision and recall Recall are greater than the screening threshold based on the screening threshold.

[0073] See also Figure 2D, showing a schematic diagram of the principle of semantic filtering of the original surface quality judgment standard.

[0074] Optionally, in this embodiment, there are multiple original surface quality judgment standards, and before using the verification set to verify the original surface quality judgment standard to obtain the original surface quality judgment standard corresponding to the surface defect category, the method also includes: using a tree model to divide the defect characteristics and defect thresholds included in each of the multiple original surface quality judgment standards to obtain the original surface quality judgment standard corresponding to the leaf node of the tree model; for any leaf node of the tree model, if the leaf node corresponds to multiple original surface quality judgment standards, deleting some of the original surface quality judgment standards until the leaf node corresponds to one original surface quality judgment standard.

[0075] Specifically, see Figure 2D , according to the defect features included in the original surface quality discrimination standard, such as "defect quantity", "defect cumulative length", "defect maximum area" and the operators ">", "≤" of the defect thresholds corresponding to each defect feature, multiple original surface quality discrimination standards are divided. When there are multiple original surface quality discrimination standards corresponding to a leaf node of a tree model, for example, rule1 and rule2 correspond to a leaf node, rule4 and rule5 correspond to a leaf node, it indicates that the semantic similarity between rule1 and rule2 is high, and the semantic similarity between rule4 and rule5 is high, then some of the original surface quality discrimination standards can be deleted until the leaf node corresponds to one original surface quality discrimination standard.

[0076] Specifically, for multiple surface quality original discriminant criteria corresponding to the same leaf node, the prediction index of each surface quality original discriminant criterion for out-of-bag samples (OOB) can be calculated, and the surface quality original discriminant criteria with high prediction index are retained (the index here uses the F1-score which combines precision and recall).

[0077] F1-score: F1-score is an indicator used in statistics to measure the accuracy of a binary classification model. It takes into account both the precision and recall of the classification model. It is the harmonic mean of the precision and recall, with a maximum of 1 and a minimum of 0.

[0078] S208: Use the verification set to verify the original surface quality judgment standard to obtain the surface quality judgment standard corresponding to the surface defect category.

[0079] In this embodiment, the original surface quality judgment standard can be verified based on the verification set to obtain indicators such as accuracy, recall rate, F1, etc. corresponding to the original surface quality judgment standard, and the surface quality judgment standard corresponding to the surface defect category can be obtained according to the corresponding indicators.

[0080] For example, feature vectors or feature codes may be used in the above processing process. In this embodiment, the surface quality judgment standard corresponding to the obtained surface defect category can also be converted into natural language to obtain a human-readable judgment standard, and the judgment standard can be applied after inspection by relevant technical personnel.

[0081] In addition, when there are multiple surface defect categories, a corresponding surface quality judgment standard can be generated for each surface defect category. There can be multiple surface quality judgment standards corresponding to each surface defect category, which is not limited in this embodiment.

[0082] The method for generating the surface quality judgment standard of this embodiment can be executed by any appropriate electronic device with data processing capability, including but not limited to: a server, a mobile terminal (such as a mobile phone, a PAD, etc.) and a PC, etc.

[0083] Figure 3 A flow chart of a surface quality inspection method of a target object provided in an embodiment of the present application is shown in the figure, which includes:

[0084] S301, obtaining defect feature information of the target object in each surface defect category;

[0085] S302 , based on the surface quality judgment standard corresponding to each surface defect category, and according to the defect feature information of the target object in each surface defect category, determine whether the target object meets the preset standard.

[0086] The surface quality criterion is determined according to the method for generating the surface quality criterion provided in the above embodiment.

[0087] Surface quality inspection: referred to as surface inspection, refers to the quality inspection of the surface of steel coils, aluminum coils, cloth and other products. The inspection content mainly targets surface defects, and conducts inspection and identification to ultimately achieve surface quality judgment and grading.

[0088] It should be noted that when there are multiple surface defect categories, the defect feature information of the target object in each surface defect category must meet the surface quality judgment standard of the corresponding surface defect category, otherwise the target object is judged to be unqualified. However, if there can be multiple surface quality judgment standards corresponding to any surface defect category, the target object only needs to meet one of them.

[0089] The surface quality inspection method of the target object of this embodiment can be executed by any appropriate electronic device with data processing capabilities, including but not limited to: a server, a mobile terminal (such as a mobile phone, a PAD, etc.) and a PC, etc.

[0090] Figure 4 A schematic flow chart of a steel coil surface quality inspection method provided in an embodiment of the present application, as shown in the figure, includes:

[0091] S401, obtaining defect feature information of the target steel coil in each surface defect category;

[0092] S402, based on the surface quality judgment standard corresponding to each surface defect category, according to the defect feature information of the target steel coil in each surface defect category, determine whether the target steel coil meets the preset standard.

[0093] The surface quality criterion is determined according to the method for generating the surface quality criterion provided in the above embodiment.

[0094] The steel coil surface quality inspection method of this embodiment can be executed by any appropriate electronic device with data processing capabilities, including but not limited to: servers, mobile terminals (such as mobile phones, PADs, etc.) and PCs, etc.

[0095] Reference Figure 5 , shows a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The specific embodiment of the present application does not limit the specific implementation of the electronic device.

[0096] like Figure 5 As shown, the electronic device may include: a processor (processor) 502 , a communication interface (Communications Interface) 504 , a memory (memory) 506 , and a communication bus 508 .

[0097] in:

[0098] The processor 502 , the communication interface 504 , and the memory 506 communicate with each other via a communication bus 508 .

[0099] The communication interface 504 is used to communicate with other electronic devices or servers.

[0100] The processor 502 is used to execute the program 510, and specifically can execute the relevant steps in the above-mentioned surface quality judgment standard generation method, target object surface quality inspection method or steel coil surface quality inspection method embodiment.

[0101] Specifically, the program 510 may include program codes, which include computer operation instructions.

[0102] The processor 502 may be a processor CPU, or an application specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application. The one or more processors included in the smart device may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0103] The memory 506 is used to store the program 510. The memory 506 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0104] The specific implementation of each step in program 510 can refer to the corresponding descriptions in the corresponding steps and units in the above-mentioned surface quality judgment standard generation method, target object surface quality inspection method or steel coil surface quality inspection method embodiment, which will not be repeated here. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the above-described equipment and modules can refer to the corresponding process description in the above-mentioned method embodiment, which will not be repeated here.

[0105] An embodiment of the present application also provides a computer program product, including computer instructions, which instruct a computing device to execute operations corresponding to any surface quality judgment standard generation method, target object surface quality inspection method or steel coil surface quality inspection method in the above-mentioned multiple method embodiments.

[0106] It should be pointed out that, according to the needs of implementation, the various components / steps described in the embodiments of the present application can be split into more components / steps, or two or more components / steps or partial operations of components / steps can be combined into new components / steps to achieve the purpose of the embodiments of the present application.

[0107] The above-mentioned method according to the embodiment of the present application can be implemented in hardware, firmware, or implemented as software or computer code that can be stored in a recording medium (such as a CD ROM, RAM, floppy disk, hard disk or magneto-optical disk), or implemented as a computer code originally stored in a remote recording medium or a non-temporary machine-readable medium downloaded through a network and to be stored in a local recording medium, so that the method described herein can be stored in such software processing on a recording medium using a general-purpose computer, a dedicated processor or programmable or dedicated hardware (such as an ASIC or FPGA). It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component (e.g., RAM, ROM, flash memory, etc.) that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the method for generating a surface quality judgment standard, a target object surface quality inspection method or a steel coil surface quality inspection method described herein is implemented. In addition, when a general-purpose computer accesses code for implementing the method for generating surface quality judgment standards, the target object surface quality inspection method, or the steel coil surface quality inspection method shown herein, the execution of the code converts the general-purpose computer into a special-purpose computer for executing the method for generating surface quality judgment standards, the target object surface quality inspection method, or the steel coil surface quality inspection method shown herein.

[0108] Those of ordinary skill in the art will appreciate that the units and method steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the embodiments of the present application.

[0109] The above implementation methods are only used to illustrate the embodiments of the present application, and are not limitations on the embodiments of the present application. Ordinary technicians in the relevant technical field can make various changes and modifications without departing from the spirit and scope of the embodiments of the present application. Therefore, all equivalent technical solutions also belong to the scope of the embodiments of the present application. The scope of patent protection of the embodiments of the present application should be limited by the claims.

Claims

1. A method for generating a surface quality judgment standard, comprising: Obtaining surface defects contained in the surfaces of a plurality of sample objects, and sample labels corresponding to the sample objects, wherein the sample labels are used to characterize whether the sample objects meet preset standards; wherein the preset standards are product surface quality standards required by users who purchase the sample objects; Based on a predetermined surface defect category, and according to the surface defects contained on the surface of the sample object, obtaining defect feature information of each of the sample objects in the surface defect category; Establishing a sample data set corresponding to the surface defect category according to the defect feature information of each sample object in the surface defect category and the sample label of each sample object; Rule learning is performed based on the sample data set corresponding to the surface defect category to obtain the surface quality judgment standard corresponding to the surface defect category.

2. The method according to claim 1, wherein: The method further comprises: Determine the sample labels of the newly added sample objects according to the sample objects added within a preset time period and the feedback information of the users who purchased the sample objects; Establishing an updated sample data set corresponding to the surface defect category according to the defect feature information of the sample objects newly added within a preset time period in the surface defect category and the sample labels; Rule learning is performed according to the updated sample data set corresponding to the surface defect category, so as to update the surface quality judgment standard corresponding to the surface defect category according to the learning result.

3. The method according to claim 1, wherein: The sample object is a steel coil or an aluminum material, the defect feature corresponding to the surface defect category includes at least one of the following, and the surface defect information is a feature value corresponding to each of the defect features: Number of defects, number of defects per meter, number of defects excluding head and tail, cumulative length, cumulative length ratio, cumulative width, cumulative area ratio, maximum defect length, maximum defect width, maximum defect area, maximum distance from the edge, maximum distance from the head, maximum distance from the tail.

4. The method according to claim 1, wherein: The surface defect category is determined by the following steps: Clustering the surface defects according to the morphological features of the surface defects respectively contained in the surfaces of the plurality of sample objects; The surface defect category is determined according to the clustering result.

5. The method according to claim 1, wherein: Performing rule learning according to the sample data set corresponding to the surface defect category to obtain the surface quality judgment standard corresponding to the surface defect category includes: Dividing the sample data set corresponding to the surface defect category into a training set and a validation set; Based on the random forest algorithm, rule learning is performed according to the training set to obtain the original surface quality discrimination standard corresponding to the surface defect category; The verification set is used to verify the original surface quality discrimination standard to obtain the original surface quality discrimination standard corresponding to the surface defect category.

6. The method according to claim 5, wherein: The random forest algorithm is used to perform rule learning according to the training set to obtain the original surface quality discrimination standard corresponding to the surface defect category, which includes: Randomly extracting defect features from each defect feature corresponding to the surface defect category; Performing sampling with replacement on the training set, and determining a feature threshold corresponding to the defect feature according to the sampling result; Perform node splitting according to the defect feature and the feature threshold to generate a decision tree; A random forest is established based on the generated decision tree, and the original surface quality discrimination standard corresponding to the surface defect category is determined based on the defect features and feature thresholds corresponding to the leaf nodes in the random forest.

7. The method according to claim 5, wherein: The number of the original surface quality discrimination criteria is multiple, and before the original surface quality discrimination criteria are verified by using the verification set to obtain the original surface quality discrimination criteria corresponding to the surface defect category, the method further includes: Using a tree model, dividing according to the defect characteristics and defect thresholds respectively included in a plurality of the original surface quality discrimination criteria, to obtain the original surface quality discrimination criteria corresponding to the leaf nodes of the tree model; For any leaf node of the tree model, if the leaf node corresponds to a plurality of the original surface quality criteria, some of the original surface quality criteria are deleted until the leaf node corresponds to one original surface quality criteria.

8. A method for inspecting the surface quality of a target object, comprising: Obtain defect feature information of the target object in each surface defect category; Based on the surface quality judgment standard corresponding to each surface defect category, it is judged whether the target object meets the preset standard according to the defect feature information of the target object in each surface defect category, and the surface quality judgment standard is determined according to the method described in claims 1-7.

9. A method for inspecting the surface quality of a steel coil, comprising: Obtain defect feature information of the target steel coil in each surface defect category; Based on the surface quality judgment standard corresponding to each surface defect category, it is judged whether the target steel coil meets the preset standard according to the defect feature information of the target steel coil in each surface defect category, and the surface quality judgment standard is determined according to the method described in claims 1-7.

10. A computer storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method according to any one of claims 1 to 9.

11. A computer program product, comprising computer instructions, wherein the computer instructions instruct a computing device to execute the method according to any one of claims 1 to 9.

Images