Method, device, storage medium and processor for detecting weld quality

Abstract

An embodiment of the present application provides a method, a device, a processor, and a storage medium for detecting the quality of a weld seam. The method includes: obtaining a similarity mean value corresponding to the information feature included in each piece of information to be detected, where each piece of information to be detected is determined after cutting the information to be detected, and the information to be detected is generated during the welding of a target article; inputting the similarity mean value into a quality detection model to determine the type of welding quality during the welding of the target article through the quality detection model according to the input similarity mean value; where the training data of the quality detection model is determined according to the feature similarity mean value between the historical information features included in each welding historical process information and the features included in the candidate feature set. Through the above technical solution, the detection efficiency of weld seam quality detection can be improved. Using the trained quality detection model to detect the weld seam quality, the accuracy is relatively high, and the required labor cost is relatively high.

Description

Technical Field

[0001] The present application relates to the technical field of quality inspection, and particularly to a method, device, storage medium and processor for detecting the quality of welds. Background Art

[0002] In the current prior art, common weld quality inspection methods include magnetic particle inspection, ultrasonic inspection, eddy current inspection, etc. By using these methods to detect weld quality, the accuracy is relatively low, the timeliness is low, the labor cost required is relatively high, and the position where the weld quality problem occurs cannot be located in time.

[0003] Welding, as a basic material manufacturing process, is indispensable in the forming of large parts in the manufacturing industry. The quality of welded products is often related to the quality of welds. If serious weld quality problems occur in welded products, the load-bearing capacity of the welded product structure may be weakened, and partial fracture of the welded product structure may occur. If the weld quality problem is not detected in time at this time, major accidents may be caused. Therefore, it is particularly important to detect weld quality problems early. Summary of the Invention

[0004] The purpose of the embodiments of the present application is to provide a method, device, storage medium and processor for detecting the quality of welds.

[0005] To achieve the above purpose, the first aspect of the present application provides a method for detecting the quality of welds, including:

[0006] Obtaining the average similarity corresponding to the information feature included in each piece of information to be detected, where each piece of information to be detected is determined after cutting the information to be detected, and the information to be detected is generated during the welding of the target article;

[0007] Inputting the average similarity into a quality inspection model, so as to determine the welding quality type during the welding of the target article according to the input average similarity by the quality inspection model;

[0008] Among them, the training data of the quality inspection model is determined according to the average feature similarity between the historical information features included in each welding historical process information and the features included in the candidate feature set. Each historical information feature is determined after cutting the welding historical process information in chronological order, and the candidate feature set is determined after screening the historical information features.

[0009] Optionally, the method further includes a training step for the quality detection model, including: obtaining a plurality of welding history data, where the welding history data includes welding history process information; cutting the welding history process information in chronological order to obtain a plurality of historical information features; screening the historical information features to obtain a candidate feature set; determining the average feature similarity between the historical information features included in each welding history process information and the features included in the candidate feature set; and inputting the average feature similarity into the quality detection model to train the quality detection model.

[0010] Optionally, cutting the welding history process information in chronological order to obtain a plurality of historical information features includes: extracting features from the welding history process information to obtain historical information features corresponding to the welding history process information; sorting the historical information features according to the time tags of each historical information feature; and cutting the welding history process information in chronological order to determine the historical information features corresponding to each welding history process information.

[0011] Optionally, each historical information feature includes a plurality of feature points. Screening the historical information features to obtain a candidate feature set includes: for each historical information feature, selecting any one feature point from the plurality of feature points included in the historical information feature as the first feature point; determining a second feature point corresponding to the first feature point according to a preset information length; using the plurality of feature points between the first feature point and the second feature point as candidate feature points; and determining the information feature composed of the candidate feature points as the target information feature to obtain a candidate feature set constructed by the target information features.

[0012] Optionally, determining the average feature similarity between the historical information features included in each welding history process information and the features included in the candidate feature set includes: determining the welding quality type corresponding to each welding history process information; classifying each welding history process information according to the welding quality type; for each welding quality type, determining the feature similarity between each historical information feature included in the welding history process information corresponding to the welding quality type and each feature included in the candidate feature set; and determining the average feature similarity between the historical information features included in each welding history process information and the features included in the candidate feature set according to all the feature similarities.

[0013] Optionally, the method further includes: after determining the average value of the feature similarities between the historical information features included in each welding history process information and the features included in the candidate feature set according to all the feature similarities, arranging the average values of the feature similarities corresponding to the multiple welding history process information included in each welding quality type in descending order to obtain an evaluation matrix of the average values of the feature similarities; screening a preset number of average values of the feature similarities from the average values of the feature similarities corresponding to each welding quality type in the evaluation matrix in descending order as the target average values of the feature similarities; and inputting the target average values of the feature similarities into the quality detection model to train the quality detection model.

[0014] Optionally, the evaluation matrix is determined by formula (1):

[0015]

[0016] where S represents the evaluation matrix of the average values of the feature similarities, and a, b, c, and d respectively correspond to different welding quality types. a n represents the average value of the feature similarities between each historical information feature included in the welding history process information corresponding to the a welding quality type and the nth target information feature included in the candidate feature set, b n represents the average value of the feature similarities between each historical information feature included in the welding history process information corresponding to the b welding quality type and the nth target information feature included in the candidate feature set, c n represents the average value of the feature similarities between each historical information feature included in the welding history process information corresponding to the c welding quality type and the nth target information feature included in the candidate feature set, and d n represents the average value of the feature similarities between each historical information feature included in the welding history process information corresponding to the d welding quality type and the nth target information feature included in the candidate feature set.

[0017] The second aspect of the present application provides a processor configured to execute the above method for detecting the quality of a weld seam.

[0018] The third aspect of the present application provides a device for detecting the quality of a weld seam, including the above processor.

[0019] The fourth aspect of the present application provides a machine-readable storage medium, on which instructions are stored, and when the instructions are executed by the processor, the processor is configured to execute the above method for detecting the quality of a weld seam.

[0020] Through the above technical solutions, the detection efficiency of weld seam quality detection can be improved. Using the trained quality detection model to detect the quality of a weld seam has relatively high accuracy and requires a relatively high labor cost.

[0021] Other features and advantages of the embodiments of the present application will be described in detail in the following specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The drawings are used to provide a further understanding of the embodiments of the present application, and constitute a part of the specification. Together with the following specific implementation, they are used to explain the embodiments of the present application, but do not constitute a limitation to the embodiments of the present application. In the drawings:

[0023] Figure 1 A schematic flowchart of a method for detecting the quality of a weld seam according to an embodiment of the present application is schematically shown;

[0024] Figure 2 The internal structure diagram of a computer device according to an embodiment of the present application is schematically shown. SPECIFIC IMPLEMENTATION

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. It should be understood that the specific implementation described here is only used to explain and illustrate the embodiments of the present application, and is not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present application.

[0026] Figure 1 A schematic flowchart of a method for detecting the quality of a weld seam according to an embodiment of the present application is schematically shown. As Figure 1 shown, in an embodiment of the present application, a method for detecting the quality of a weld seam is provided, including the following steps:

[0027] Step 101: Obtain the average similarity corresponding to the information features included in each piece of information to be detected, where each piece of information to be detected is determined after cutting the information to be detected, and the information to be detected is generated during the welding of the target article.

[0028] Step 102: Input the average similarity into a quality detection model to determine the welding quality type during the welding of the target article according to the input average similarity by the quality detection model.

[0029] Among them, the training data of the quality detection model is determined according to the average feature similarity between the historical information features included in each welding historical process information and the features included in the candidate feature set. Each historical information feature is determined after cutting the welding historical process information in chronological order, and the candidate feature set is determined by screening the historical information features.

[0030] To determine the type of welding quality when welding a target item, the processor can first obtain the average similarity corresponding to the information features included in each piece of information to be detected. Specifically, the processor can obtain the information to be detected during the welding of the target item. The target item can refer to the product that needs to be welded. The information to be detected can refer to the welding process information. The welding process information can include high-frequency real-time information of welding, set parameters of the welding robot, and external parameters affecting the welding quality, etc. Specifically, the high-frequency real-time information of welding can include the welding attitude inclination angle, real-time welding current, real-time welding voltage, molten pool image, and welding sound, etc. The set parameters of the welding robot can include welding speed, wire feeding speed of the robot, welding voltage, and welding current, etc. The external parameters affecting the welding quality can include the type of welding material and the thickness of the welding material, etc. The processor can cut the information to be detected in chronological order to obtain multiple information features included in each piece of information to be detected. Then, the processor can screen each information feature and determine the average similarity between the information features included in each piece of information to be detected and the screened information features.

[0031] The processor can input the average similarity into the trained quality detection model to determine the type of welding quality when welding the target item through the quality detection model according to the input average similarity. Among them, the training data of the quality detection model is determined according to the average feature similarity between the historical information features included in each welding historical process information and the features included in the candidate feature set. Each historical information feature is determined after cutting the welding historical process information in chronological order. The candidate feature set is determined after screening the historical information features.

[0032] Through the above technical solution, the detection efficiency of weld quality detection can be improved. Using the trained quality detection model to detect the weld quality has relatively high accuracy, but requires a relatively high labor cost.

[0033] In one embodiment, the method further includes a training step for the quality detection model, including: obtaining a plurality of welding historical data, where the welding historical data includes welding historical process information; cutting the welding historical process information in chronological order to obtain a plurality of historical information features; screening the historical information features to obtain a candidate feature set; determining the average feature similarity between the historical information features included in each welding historical process information and the features included in the candidate feature set; inputting the average feature similarity into the quality detection model to train the quality detection model.

[0034] To train a quality inspection model, a processor can obtain multiple welding history data. The welding history data can include welding history process information. Among them, the welding history process information can include historical welding high-frequency information, welding robot setting parameters, and external parameters affecting welding quality, etc. Specifically, the historical welding high-frequency information can include historical welding attitude inclination, historical welding current, historical welding voltage, historical molten pool image, and historical welding sound, etc. The processor can cut the welding history process information in chronological order to obtain multiple historical information features. Then, the processor can screen the historical information features to obtain a candidate feature set. The processor can determine the average value of the feature similarity between the historical information features included in each welding history process information and the features included in the candidate feature set, and input the average value of the feature similarity into the quality inspection model to train the quality inspection model.

[0035] In one embodiment, cutting the welding history process information in chronological order to obtain multiple historical information features includes: extracting features from the welding history process information to obtain historical information features corresponding to the welding history process information; sorting the historical information features according to the time tags of each historical information feature; and cutting the welding history process information according to chronological order to determine the historical information features corresponding to each welding history process information.

[0036] The processor can extract features from the welding history process information to obtain historical information features corresponding to the welding history process information. If the welding history process information includes a historical molten pool image, the corresponding historical information features can be molten pool width, molten pool length, molten pool perimeter, and time tag, etc. If the welding history process information includes historical welding sound, the corresponding historical information features can be sound pressure, frequency spectrum, and time tag, etc. The processor can sort the historical information features according to the time tags of each historical information feature and cut the welding history process information according to chronological order to determine the historical information features corresponding to each welding history process information.

[0037] For example, if the welding history process information is M, after the processor extracts the features of the welding history process information, it can obtain the historical information features A, B, and C corresponding to the welding history process information. Each historical information feature may include a time tag. The processor can sort the historical information features A, B, and C according to the time tag to correspond the historical information features at the same time point. Then, the processor can cut the welding history process information M according to the time sequence to obtain multiple welding history process information M1, M2, M3, and M4. After the welding history process information M is cut, its corresponding historical information features A, B, and C are also cut accordingly. That is, the welding history process information M1 may include the historical information features A1, B1, and C1, the welding history process information M2 may include the historical information features A2, B2, and C2, the welding history process information M3 may include the historical information features A3, B3, and C3, and the welding history process information M4 may include the historical information features A4, B4, and C4.

[0038] In one embodiment, the welding history process information is cut according to the weld length. First, the processor can extract the features of the welding history process information of the weld to obtain the historical information features corresponding to the welding history process information. Then, the processor can cut the welding history process information of the weld according to the weld length to determine the historical information features included in the welding history process information corresponding to each weld. For example, if the length of the weld is 10 mm, the welding history process information of the weld can be cut according to the weld length of every 1 mm to determine the historical information features included in the welding history process information corresponding to each 1 mm weld after cutting.

[0039] In one embodiment, each historical information feature contains multiple feature points. The historical information features are screened to obtain a candidate feature set, including: for each historical information feature, selecting any one feature point from the multiple feature points included in the historical information feature as the first feature point; determining a second feature point corresponding to the first feature point according to the preset information length; taking the multiple feature points included between the first feature point and the second feature point as candidate feature points; and determining the information feature composed of the candidate feature points as the target information feature to obtain a candidate feature set constructed by the target information features.

[0040] For each historical information feature, the processor may select any one feature point from the multiple feature points included in the historical information feature as the first feature point. Then, the processor may determine a second feature point corresponding to the first feature point according to a preset information length. The processor may use the multiple feature points included between the first feature point and the second feature point as candidate feature points. The processor may determine the information feature constituted by the candidate feature points as the target information feature, so as to obtain a candidate feature set constructed by the target information feature. Among them, the weld quality type can be distinguished to the greatest extent through the target information features included in the candidate feature set.

[0041] For example, the historical information feature A1 includes 100 feature points, and the preset information length is the information length constituted by 80 feature points. For the historical information feature A1 included in the welding historical process information M1, the processor may select any one feature point from the 100 feature points included in the historical information feature A1 as the first feature point. The processor may determine a second feature point corresponding to the first feature point according to the preset information length. If the processor uses the first information feature point of the historical information feature A1 as the first feature point, then according to the preset information length, the second feature point can be determined as the 80th information feature point of the historical information feature A1. Then, the processor may use all the feature points included between the first information feature point and the 80th information feature point of the historical information feature A1 as candidate feature points. The processor may select other information feature points in the historical information feature A1 as the first feature point until all the second feature points corresponding to the first feature point are determined according to the preset information length, so as to determine multiple target information features. The processor may construct a candidate feature set according to the multiple target information features.

[0042] In one embodiment, determining the average feature similarity between the historical information features included in each welding historical process information and the features included in the candidate feature set includes: determining the welding quality type corresponding to each welding historical process information; classifying each welding historical process information according to the welding quality type; for each welding quality type, determining the feature similarity between each historical information feature included in the welding historical process information corresponding to the welding quality type and each feature included in the candidate feature set; and determining the average feature similarity between the historical information features included in each welding historical process information and the features included in the candidate feature set according to all the feature similarities.

[0043] The processor can determine the welding quality type corresponding to each welding history process information. Among them, the welding quality type can be the presence of pores in the weld seam, no pores in the weld seam, the presence of weld beads after welding, no weld beads after welding, and weld deviation, etc. The presence of pores in the weld seam may be caused by the environment or improper operation. Weld deviation may be caused by inaccurate welding position. The presence of weld beads after welding may be caused by incorrect process parameters or poor operation. The processor can classify each welding history process information according to the welding quality type. For each welding quality type, the processor can determine the feature similarity between each historical information feature included in the welding history process information corresponding to the welding quality type and each feature included in the candidate feature set. The process can determine the average feature similarity between the historical information features included in each welding history process information and the features included in the candidate feature set according to all the feature similarities.

[0044] For example, if the welding history process information M is divided to obtain welding history process information M1, M2, M3, and M4. The processor can determine that the welding quality types corresponding to the welding history process information M1, M2, M3, and M4 are X, Y, X, and Y respectively. That is, it can be shown that the welding quality types corresponding to the welding history process information M1 and M3 are the same, and the welding quality types corresponding to the welding history process information M2 and M4 are the same. Therefore, the processor can classify the welding history process information M1 and M3 into one category according to the welding quality type X, and classify the welding history process information M2 and M4 into one category according to the welding quality type Y. For the welding quality type corresponding to the welding history process information M1 and M3, the processor can determine the historical information features A1, B1, C1 included in the welding history process information M1 and the historical information features A3, B3, C3 included in the welding history process information M3, and the feature similarities with the first feature included in the candidate feature set respectively. In the case of determining all the feature similarities, the processor can determine the average feature similarity a1 between the historical information features included in each welding history process information and the features included in the candidate feature set according to all the feature similarities. For example, the processor can determine the average feature similarity a1 by averaging all the feature similarities. After that, the processor can determine the historical information features A1, B1, C1 included in the welding history process information M1 and the historical information features A3, B3, C3 included in the welding history process information M3, and the feature similarities with the second feature included in the candidate feature set respectively, to determine the average feature similarity a2, until the feature similarities with all the features included in the candidate feature set are determined, and all the corresponding average feature similarities are determined.

[0045] In one embodiment, the method further includes: after determining the average feature similarity between the historical information features included in each welding history process information and the features included in the candidate feature set according to all the feature similarities, arranging the average feature similarities corresponding to the multiple welding history process information included in each welding quality type in descending order to obtain an evaluation matrix of the average feature similarities; screening out a preset number of average feature similarities as target average feature similarities from the average feature similarities corresponding to each welding quality type in the evaluation matrix in descending order; and inputting the target average feature similarities into the quality detection model to train the quality detection model.

[0046] After determining the average feature similarity between the historical information features included in each welding history process information and the features included in the candidate feature set according to all the feature similarities, the processor may arrange the average feature similarities corresponding to the multiple welding history process information included in each welding quality type in descending order to obtain an evaluation matrix of the average feature similarities. In descending order, the processor may screen out a preset number of average feature similarities from the average feature similarities corresponding to each welding quality type in the evaluation matrix as target average feature similarities. For example, if the average feature similarities corresponding to each welding quality type include n, then the processor may select m average feature similarities in descending order as target average feature similarities, where n ≥ m. Then, the processor may input the target average feature similarities into the quality detection model to train the quality detection model. Among them, the quality detection model may be a machine learning classification model, such as xgboost, etc.

[0047] In one embodiment, the evaluation matrix is determined by formula (1):

[0048]

[0049] where S represents the evaluation matrix of the average feature similarities, and a, b, c, and d respectively correspond to different welding quality types. a n represents the average feature similarity between each historical information feature included in the welding history process information corresponding to the a welding quality type and the nth target information feature included in the candidate feature set. b n represents the average feature similarity between each historical information feature included in the welding history process information corresponding to the b welding quality type and the nth target information feature included in the candidate feature set. c n represents the average feature similarity between each historical information feature included in the welding history process information corresponding to the c welding quality type and the nth target information feature included in the candidate feature set. d nIt represents the average feature similarity between each historical information feature included in the welding historical process information corresponding to the welding quality type denoted as d and the nth target information feature included in the candidate feature set.

[0050] Among them, the welding quality type can be the presence of pores in the weld, no pores in the weld, the presence of weld beads after welding, no weld beads after welding, and welding deviation, etc. The average feature similarities corresponding to the multiple welding historical process information included in each welding quality type are arranged in descending order. For example, for the welding quality type a, its corresponding average feature similarities a1, a2...a n Are arranged in descending order.

[0051] Through the above technical solution, the detection efficiency of weld quality detection can be improved. Using the trained quality detection model to detect weld quality has relatively high accuracy and requires low labor costs. At the same time, by using the average feature similarity between the historical information features included in each welding historical process information and the features included in the candidate feature set as the input of the quality detection model, when detecting weld quality subsequently, the specific location where weld defects occur can be quickly located, which can reduce the time of manual quality inspection and greatly improve the production efficiency of welded products.

[0052] Figure 1 It is a schematic flowchart of a method for detecting weld quality in an embodiment. It should be understood that although Figure 1 The steps in the flowchart are shown in sequence according to the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, Figure 1 At least a part of the steps in

[0053] This application embodiment provides a processor. The processor is used to run a program. Among them, when the program runs, it executes the above method for detecting weld quality.

[0054] This application embodiment provides a device for detecting weld quality, including the above-mentioned processor.

[0055] This application embodiment provides a storage medium, on which a program is stored. When the program is executed by the processor, it implements the above method for detecting weld quality.

[0056] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structural diagram may be as shown in Figure 2 . The computer device includes a processor A01, a network interface A02, a memory (not shown in the figure), and a database (not shown in the figure) connected through a system bus. Among them, the processor A01 of the computer device is used to provide computing and control capabilities. The memory of the computer device includes an internal memory A03 and a non-volatile storage medium A04. The non-volatile storage medium A04 stores an operating system B01, a computer program B02, and a database (not shown in the figure). The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 in the non-volatile storage medium A04. The database of the computer device is used to store data such as information to be detected. The network interface A02 of the computer device is used to communicate with an external terminal through a network connection. When the computer program B02 is executed by the processor A01, it implements a method for detecting the quality of a weld.

[0057] Those skilled in the art can understand that Figure 2 the structure shown in is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0058] An embodiment of the present application provides a device. The device includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, the following steps are implemented: obtaining the similarity mean value corresponding to the information feature included in each piece of information to be detected, where each piece of information to be detected is determined after cutting the information to be detected, and the information to be detected is generated during the welding of a target item; inputting the similarity mean value into a quality detection model to determine the welding quality type during the welding of the target item according to the input similarity mean value through the quality detection model; where the training data of the quality detection model is determined according to the feature similarity mean value between the historical information feature included in each welding historical process information and the feature included in the candidate feature set, each historical information feature is determined after cutting the welding historical process information in chronological order, and the candidate feature set is determined after screening the historical information features.

[0059] In one embodiment, the method further includes a training step for the quality detection model, including: obtaining a plurality of welding history data, where the welding history data includes welding history process information; cutting the welding history process information in chronological order to obtain a plurality of historical information features; screening the historical information features to obtain a candidate feature set; determining the average feature similarity between the historical information features included in each welding history process information and the features included in the candidate feature set; and inputting the average feature similarity into the quality detection model to train the quality detection model.

[0060] In one embodiment, cutting the welding history process information in chronological order to obtain a plurality of historical information features includes: extracting features from the welding history process information to obtain historical information features corresponding to the welding history process information; sorting the historical information features according to the time tags of each historical information feature; and cutting the welding history process information in chronological order to determine the historical information features corresponding to each welding history process information.

[0061] In one embodiment, each historical information feature includes a plurality of feature points. Screening the historical information features to obtain a candidate feature set includes: for each historical information feature, selecting any one feature point from the plurality of feature points included in the historical information feature as the first feature point; determining a second feature point corresponding to the first feature point according to a preset information length; taking the plurality of feature points between the first feature point and the second feature point as candidate feature points; and determining the information feature composed of the candidate feature points as the target information feature to obtain a candidate feature set constructed by the target information features.

[0062] In one embodiment, determining the average feature similarity between the historical information features included in each welding history process information and the features included in the candidate feature set includes: determining the welding quality type corresponding to each welding history process information; classifying each welding history process information according to the welding quality type; for each welding quality type, determining the feature similarity between each historical information feature included in the welding history process information corresponding to the welding quality type and each feature included in the candidate feature set; and determining the average feature similarity between the historical information features included in each welding history process information and the features included in the candidate feature set according to all the feature similarities.

[0063] In one embodiment, the method further includes: after determining the average value of the feature similarities between the historical information features included in each welding history process information and the features included in the candidate feature set according to all the feature similarities, arranging the average values of the feature similarities corresponding to the multiple welding history process information included in each welding quality type in descending order to obtain an evaluation matrix of the average values of the feature similarities; screening out a preset number of average values of the feature similarities as target similarity average values from the average values of the feature similarities corresponding to each welding quality type in the evaluation matrix in descending order; and inputting the target similarity average values into the quality detection model to train the quality detection model.

[0064] In one embodiment, the evaluation matrix is determined by formula (1):

[0065]

[0066] where S represents the evaluation matrix of the average values of the feature similarities, and a, b, c, and d respectively correspond to different welding quality types. a n represents the average value of the feature similarities between each historical information feature included in the welding history process information corresponding to the a welding quality type and the nth target information feature included in the candidate feature set, b n represents the average value of the feature similarities between each historical information feature included in the welding history process information corresponding to the b welding quality type and the nth target information feature included in the candidate feature set, c n represents the average value of the feature similarities between each historical information feature included in the welding history process information corresponding to the c welding quality type and the nth target information feature included in the candidate feature set, d n represents the average value of the feature similarities between each historical information feature included in the welding history process information corresponding to the d welding quality type and the nth target information feature included in the candidate feature set.

[0067] The present application also provides a computer program product, which, when executed on a data processing device, is adapted to execute a program initialized with the following method steps: obtaining a similarity mean value corresponding to an information feature included in each information to be detected, where each information to be detected is determined after the information to be detected is segmented, and the information to be detected is generated during the welding of a target article; inputting the similarity mean value into a quality detection model to determine, by the quality detection model according to the input similarity mean value, the welding quality type during the welding of the target article; where the training data of the quality detection model is determined according to the feature similarity mean value between the historical information feature included in each welding historical process information and the features included in the candidate feature set, each historical information feature is determined after the welding historical process information is segmented in chronological order, and the candidate feature set is determined after screening the historical information features.

[0068] In one embodiment, the method further includes a training step for the quality detection model, including: obtaining a plurality of welding historical data, where the welding historical data includes welding historical process information; segmenting the welding historical process information in chronological order to obtain a plurality of historical information features; screening the historical information features to obtain a candidate feature set; determining the feature similarity mean value between the historical information feature included in each welding historical process information and the features included in the candidate feature set; and inputting the feature similarity mean value into the quality detection model to train the quality detection model.

[0069] In one embodiment, segmenting the welding historical process information in chronological order to obtain a plurality of historical information features includes: extracting features from the welding historical process information to obtain historical information features corresponding to the welding historical process information; sorting the historical information features according to the time tags of each historical information feature; and segmenting the welding historical process information in chronological order to determine the historical information features corresponding to each welding historical process information.

[0070] In one embodiment, each historical information feature includes a plurality of feature points. Screening the historical information features to obtain a candidate feature set includes: for each historical information feature, selecting any one feature point from the plurality of feature points included in the historical information feature as the first feature point; determining a second feature point corresponding to the first feature point according to a preset information length; taking the plurality of feature points included between the first feature point and the second feature point as candidate feature points; and determining the information feature composed of the candidate feature points as the target information feature to obtain a candidate feature set constructed by the target information features.

[0071] In one embodiment, determining the average feature similarity between the historical information features included in each welding history process information and the features included in the candidate feature set includes: determining the welding quality type corresponding to each welding history process information; classifying each welding history process information according to the welding quality type; for each welding quality type, determining the feature similarity between each historical information feature included in the welding history process information corresponding to the welding quality type and each feature included in the candidate feature set; and determining the average feature similarity between the historical information features included in each welding history process information and the features included in the candidate feature set according to all the feature similarities.

[0072] In one embodiment, the method further includes: after determining the average feature similarity between the historical information features included in each welding history process information and the features included in the candidate feature set according to all the feature similarities, arranging the average feature similarities corresponding to the multiple welding history process information included in each welding quality type in descending order to obtain an evaluation matrix of the average feature similarities; screening a preset number of average feature similarities from the average feature similarities corresponding to each welding quality type in the evaluation matrix in descending order as the target average feature similarities; and inputting the target average feature similarities into the quality detection model to train the quality detection model.

[0073] In one embodiment, the evaluation matrix is determined by formula (1):

[0074]

[0075] where S represents the evaluation matrix of the average feature similarities, and a, b, c, and d respectively correspond to different welding quality types. a n represents the average feature similarity between each historical information feature included in the welding history process information corresponding to the a welding quality type and the nth target information feature included in the candidate feature set, b n represents the average feature similarity between each historical information feature included in the welding history process information corresponding to the b welding quality type and the nth target information feature included in the candidate feature set, c n represents the average feature similarity between each historical information feature included in the welding history process information corresponding to the c welding quality type and the nth target information feature included in the candidate feature set, and d n represents the average feature similarity between each historical information feature included in the welding history process information corresponding to the d welding quality type and the nth target information feature included in the candidate feature set.

[0076] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0077] The present application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for realizing the functions specified in Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0078] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means realizes the functions specified in Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0079] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for realizing the functions specified in Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0080] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and a memory.

[0081] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM), and / or non-volatile memory in the form of, for example, read-only memory (ROM) or flash memory (flash RAM). The memory is an example of a computer-readable medium.

[0082] A computer-readable medium includes permanent and non-permanent, removable and non-removable media that can implement information storage by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transitory medium that can be used to store information accessible by a computing device. As defined herein, a computer-readable medium does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0083] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0084] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. A method for detecting the quality of a weld seam, characterized in that, The method includes: Obtaining the average similarity corresponding to the information features included in each piece of information to be detected, where each piece of information to be detected is determined after cutting the information to be detected, and the information to be detected is generated during the welding of the target article; Inputting the average similarity into a quality detection model, so as to determine the welding quality type during the welding of the target article according to the input average similarity by the quality detection model; Among them, the training data of the quality detection model is determined according to the average feature similarity between the historical information features included in each welding historical process information and the features included in the candidate feature set. Each historical information feature is determined after cutting the welding historical process information in chronological order, and the candidate feature set is determined after screening the historical information features; Among them, according to the average feature similarity between the historical information features included in each welding historical process information and the features included in the candidate feature set includes: Determining the welding quality type corresponding to each welding historical process information; Classifying each welding historical process information according to the welding quality type; For each welding quality type, determining the feature similarity between each historical information feature included in the welding historical process information corresponding to the welding quality type and each feature included in the candidate feature set; Determining the average feature similarity between the historical information features included in each welding historical process information and the features included in the candidate feature set according to all the feature similarities.

2. The method for detecting the quality of a weld seam according to claim 1, characterized in that, The method further includes a training step for the quality detection model, including: Obtaining a plurality of welding historical data, where the welding historical data includes welding historical process information; Cutting the welding historical process information in chronological order to obtain a plurality of historical information features; Screening the historical information features to obtain a candidate feature set; Determining the average feature similarity between the historical information features included in each welding historical process information and the features included in the candidate feature set; Inputting the average feature similarity into the quality detection model to train the quality detection model.

3. The method for detecting the quality of a weld seam according to claim 2, wherein, The cutting the welding historical process information in chronological order to obtain a plurality of historical information features includes: Performing feature extraction on the welding historical process information to obtain the historical information features corresponding to the welding historical process information; Sorting the historical information features according to the time tags of each historical information feature; Cutting the welding historical process information in chronological order to determine the historical information features corresponding to each welding historical process information.

4. The method for detecting the quality of a weld seam according to claim 2, characterized in that, Each historical information feature includes a plurality of feature points. The screening the historical information features to obtain a candidate feature set includes: For each historical information feature, selecting any one feature point from the multiple feature points included in the historical information feature as the first feature point; Determining a second feature point corresponding to the first feature point according to a preset information length; Taking the multiple feature points included between the first feature point and the second feature point as candidate feature points; Determine the information feature composed of the candidate feature points as the target information feature to obtain a candidate feature set constructed by the target information feature.

5. The method for detecting the quality of a weld seam according to claim 1, characterized in that, The method further includes: After determining the mean value of the feature similarity between the historical information features included in each welding history process information and the features included in the candidate feature set according to all the feature similarities, arrange the mean values of the feature similarities corresponding to the multiple welding history process information included in each welding quality type in descending order to obtain an evaluation matrix of the mean values of the feature similarities; Select a preset number of mean values of the feature similarities as the target mean values of the feature similarities from the mean values of the feature similarities corresponding to each welding quality type in the evaluation matrix in descending order; Input the target mean value of the feature similarity into the quality detection model to train the quality detection model.

6. The method for detecting the quality of a weld seam according to claim 5, characterized in that, The evaluation matrix is determined by formula (1): Among them, S represents the evaluation matrix of the average feature similarity. a, b, c, and d respectively correspond to different welding quality types, and a n represents the average feature similarity between each historical information feature included in the welding historical process information corresponding to the a welding quality type and the nth target information feature included in the candidate feature set. b n represents the average feature similarity between each historical information feature included in the welding historical process information corresponding to the b welding quality type and the nth target information feature included in the candidate feature set. c n represents the average feature similarity between each historical information feature included in the welding historical process information corresponding to the c welding quality type and the nth target information feature included in the candidate feature set. d n represents the average feature similarity between each historical information feature included in the welding historical process information corresponding to the d welding quality type and the nth target information feature included in the candidate feature set.

7. A processor, characterized in that, Configured to execute the method for detecting weld quality according to any one of claims 1 to 6.

8. A device for detecting the quality of a weld seam, characterized in that, Including a processor according to claim 7.

9. A machine-readable storage medium having instructions stored thereon, characterized in that, When executed by the processor, the instruction causes the processor to be configured to execute the method for detecting weld quality according to any one of claims 1 to 6.

Images