A robotic vision flexible inspection system and method

Abstract

The application discloses a kind of robot vision flexible detection system and method, it is related to welding quality detection field, including data acquisition module, image detection module, interference fringe analysis module and comprehensive optimization module, by detecting and evaluating to the interference fringe image, gray image and contour image of welding, it is favorable to evaluate welding quality immediately after welding is completed, improve the accuracy and effectiveness of welding quality detection, discover and repair potential defects in time, ensure that welding quality meets requirements, while the detection result of interference fringe and gray image can be used as important basis for quality traceability, help to analyze the cause of welding defect, take corresponding measures to improve, so as to improve the quality level of entire production process.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of welding quality detection, in particular to a robot vision flexible detection system and method. BACKGROUND

[0002] Welding quality detection is of great importance in ensuring structural safety, improving product quality, prolonging service life, meeting regulatory and standard requirements, enhancing corporate image and competitiveness, and preventing potential risks. Therefore, it is crucial to implement strict quality detection during the welding process. To improve the accuracy of welding quality detection, the existing technology has the following deficiencies:

[0003] The existing technology lacks necessary comprehensive reference data when detecting and analyzing the welding quality of the welded part, reducing the accuracy and effectiveness of the welding quality detection and analysis of the welded part. The existing technology usually only adopts a single measurement standard when detecting and analyzing the welding quality of the welded part, lacking multi-directional consideration, further reducing the accuracy and effectiveness of the welding quality detection and analysis of the welded part. SUMMARY

[0004] The purpose of the present application is to provide a robot vision flexible detection system to solve the problems raised in the background art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solution: a robot vision flexible detection system, comprising:

[0006] A data acquisition module for acquiring task data corresponding to the target welded part;

[0007] An image detection module for monitoring the images of the welded part of the target welded part to obtain the interference fringe images, gray images and contour images of each welding sub-region corresponding to the target welded part;

[0008] An interference fringe analysis module for analyzing the interference fringe images of each welding sub-region corresponding to the target welded part to obtain the welding quality reference coefficients of each welding sub-region, and evaluating the preliminary evaluation results of the welding quality corresponding to the target welded part;

[0009] A comprehensive optimization module for analyzing the gray images and contour images of each welding sub-region corresponding to the target welded part to obtain the welding quality evaluation coefficients and welding material welding use compliance coefficients of each welding sub-region, and comprehensively analyzing the preliminary evaluation results of the welding quality corresponding to the target welded part to obtain the comprehensive evaluation results of the welding quality corresponding to the target welded part.

[0010] In the preferred embodiment of the present application, the data acquisition module is implemented as follows:

[0011] Obtaining task data and welding data of the target welding piece through a welding work log;

[0012] The task data includes a welding length of the target welding piece;

[0013] The welding data refers to a type and a usage of welding material corresponding to the target welding piece.

[0014] In a preferred embodiment of the present solution, the image detection module is implemented as follows:

[0015] The welding length of the target welding piece and a preset interval length are used to divide the welding position of the target welding piece, so as to obtain each welding sub-region corresponding to the welding position of the target welding piece, denoted as each welding sub-region corresponding to the target welding piece;

[0016] The optical camera in the detection robot is used to detect the welding quality of the welding position of the target welding piece, so as to obtain an interference fringe image, a grayscale image and a contour image of each welding sub-region corresponding to the target welding piece.

[0017] In a preferred embodiment of the present solution, the interference fringe analysis module is implemented as follows:

[0018] Information extraction is performed on the interference fringe image of each welding sub-region corresponding to the target welding piece, so as to obtain interference fringe information of each welding sub-region corresponding to the target welding piece, wherein the interference fringe information includes the clarity of each interference fringe, the first interval and the second interval of each adjacent interference fringe, and the width and the crack number of each interference fringe;

[0019] A temperature detection device is used to detect the temperature of the target welding piece, so as to obtain the welding material temperature of each welding sub-region corresponding to the target welding piece;

[0020] An optical sensor is used to obtain the ambient light intensity corresponding to the target welding piece;

[0021] Each standard interference fringe data set stored in the database is extracted, wherein the selected data corresponding to the standard interference fringe data set includes the ambient light intensity, the welding material temperature and the welding material type, and the ambient light intensity corresponding to the target welding piece, the welding material temperature of each welding sub-region and the welding material type are used for screening, so as to obtain the standard interference fringe data set of each welding sub-region corresponding to the target welding piece, wherein the standard interference fringe data set includes the standard interference fringe clarity, the standard interference fringe interval, the standard interference fringe width and the standard interference fringe crack number;

[0022] Data analysis is performed on the interference fringe information of each welding sub-region corresponding to the target welding piece and the standard interference fringe data set of each welding sub-region corresponding to the target welding piece, so as to obtain the welding quality reference coefficient corresponding to each welding sub-region;

[0023] The welding quality reference coefficients corresponding to each welding sub-region are compared with the preset standard welding quality reference coefficients. If the welding quality reference coefficients corresponding to each welding sub-region are greater than the preset standard welding quality reference coefficients, it indicates that the preliminary welding quality assessment of the welding sub-region is normal. If the welding quality reference coefficients corresponding to each welding sub-region are less than or equal to the preset standard welding quality reference coefficients, it indicates that the preliminary welding quality assessment of the welding sub-region is abnormal. The welding sub-regions with normal preliminary welding quality assessments and the welding sub-regions with abnormal preliminary welding quality assessments corresponding to the target weldment are selected and recorded as the preliminary welding quality assessment results of the target weldment.

[0024] In the preferred embodiment of this solution, the specific execution method of the comprehensive optimization module is as follows:

[0025] By extracting data from the grayscale images of each welding sub-region of the target welded part, the grayscale value of each pixel in each welding sub-region is obtained. The mean, variance and standard deviation of the grayscale value of each pixel in each welding sub-region are calculated.

[0026] The grayscale values ​​of each pixel in each welding sub-region are statistically analyzed to obtain the grayscale histogram corresponding to each welding sub-region. By extracting data from the grayscale histograms corresponding to each welding sub-region, the number of peaks in the grayscale histograms corresponding to each welding sub-region is obtained.

[0027] A data extraction relationship is established between the comprehensive optimization module and the database. The data dispersion model stored in the database is extracted. The data to be filled in corresponding to the data dispersion model includes the mean, variance, standard deviation and grayscale peaks of the pixel grayscale values. Data analysis is performed on the mean, variance, standard deviation and grayscale peaks of the pixel grayscale values ​​in each welding sub-region and the data dispersion model to obtain the grayscale dispersion of the pixel grayscale values ​​in each welding sub-region. The grayscale dispersion of the pixel grayscale values ​​in each welding sub-region is recorded as the welding quality evaluation coefficient corresponding to each welding sub-region.

[0028] Extract the standard contour model corresponding to the target welded part stored in the database. Divide the standard contour model corresponding to the target welded part according to each welding sub-region to obtain the standard contour model corresponding to each welding sub-region. Record the standard contour model corresponding to each welding sub-region as the initial contour model corresponding to each welding sub-region.

[0029] The contour images of each welding sub-region are scanned from all directions by a contour scanner to obtain the three-dimensional images corresponding to each welding sub-region, and a post-welding contour model corresponding to each welding sub-region is established.

[0030] Compare the initial contour model corresponding to each welding sub-region with the post-welding contour model corresponding to each welding sub-region to obtain the differential contour model before and after welding for each welding sub-region, obtain the actual volume corresponding to the differential contour model, and statistically obtain the actual volume corresponding to the differential contour model of each welding sub-region.

[0031] Extract the shrinkage rates corresponding to various welding materials stored in the database, and obtain the shrinkage rate of the welding material corresponding to the target welding part by screening according to the type of welding material of the target welding part.

[0032] Obtain the used volume corresponding to the welding material based on the usage amount of the welding material.

[0033] Perform model analysis based on the shrinkage rate of the welding material corresponding to the target welding part, the used volume of the welding material corresponding to the target welding part, and the actual volume corresponding to the differential contour model of each welding sub-region to obtain the welding usage compliance coefficient corresponding to each welding sub-region.

[0034] Perform data analysis on the welding usage compliance coefficient corresponding to each welding sub-region, the welding quality evaluation coefficient corresponding to each welding sub-region, and the welding quality reference coefficient corresponding to each welding sub-region to obtain the effective welding quality evaluation coefficient corresponding to each welding sub-region. Among them, compare the effective welding quality evaluation coefficient corresponding to each welding sub-region with the preset threshold of the effective welding quality evaluation coefficient. If the effective welding quality evaluation coefficient corresponding to each welding sub-region is greater than or equal to the preset threshold of the effective welding quality evaluation coefficient, it means that the welding quality of the target welding part is qualified, and record the qualified welding quality of the target welding part as the comprehensive evaluation result of the welding quality corresponding to the target welding part. If there is a welding sub-region where the effective welding quality evaluation coefficient is less than the preset threshold of the effective welding quality evaluation coefficient, it means that the welding quality of the target welding part is unqualified. Record the welding sub-region with an effective welding quality evaluation coefficient less than the preset threshold as each abnormal sub-region, statistically obtain each abnormal sub-region corresponding to the target welding part, and record each abnormal sub-region corresponding to the target welding part as the comprehensive evaluation result of the welding quality corresponding to the target welding part.

[0035] To achieve the above object, the present invention also provides the following technical solution: A robot vision flexible detection method, including the following steps:

[0036] Obtain the task data corresponding to the target welding part.

[0037] Monitor the image of the welding area of the target welding part to obtain the interference fringe image, grayscale image, and contour image of each welding sub-region corresponding to the target welding part.

[0038] The interference fringe images of each welding sub-region of the target welded part are analyzed to obtain the welding quality reference coefficients corresponding to each welding sub-region, and the preliminary evaluation results of the welding quality of the target welded part are obtained.

[0039] The grayscale and contour images of each welding sub-region corresponding to the target welded part are analyzed to obtain the welding quality evaluation coefficient and welding material compliance coefficient corresponding to each welding sub-region. Combined with the preliminary welding quality assessment results of the target welded part, a comprehensive analysis is performed to obtain the comprehensive welding quality assessment result of the target welded part.

[0040] Compared with the prior art, the beneficial effects of the present invention are:

[0041] This invention, by detecting and evaluating interference fringe images, grayscale images, and contour images of the weld joint, facilitates immediate evaluation of weld quality after welding, improves the accuracy and effectiveness of weld quality inspection, promptly identifies and repairs potential defects, and ensures that weld quality meets requirements. Furthermore, the detection results of interference fringes and grayscale images can serve as important evidence for quality traceability, helping to analyze the causes of welding defects and take corresponding measures for improvement, thereby enhancing the overall quality level of the production process. Attached Figure Description

[0042] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0043] Figure 1 This is a schematic diagram of module connections in an embodiment of the present invention.

[0044] Figure 2 This is a schematic diagram illustrating the connection steps in an embodiment of the present invention. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] Please see Figure 1 The present invention provides a robot vision flexible inspection system, which includes a data acquisition module, an image detection module, an interference fringe analysis module and a comprehensive optimization module;

[0047] The data acquisition module is connected to the interference fringe analysis module, the image detection module is connected to the interference fringe analysis module, and the interference fringe analysis module is connected to the comprehensive optimization module.

[0048] The data acquisition module is used to acquire the task data corresponding to the target weldment.

[0049] Furthermore, the specific execution method of the data acquisition module is as follows:

[0050] Obtain task data and welding data of the target weldment through the welding work log;

[0051] The task data includes the required welding length of the target weldment;

[0052] The welding data refers to the type and amount of welding material used for the target weldment.

[0053] The image detection module is used to monitor the image of the welded area of ​​the target weldment and obtain the interference fringe image, grayscale image and contour image of each welded sub-region of the target weldment;

[0054] Furthermore, the specific execution method of the image detection module is as follows:

[0055] The welding area of ​​the target weldment is divided according to the welding length of the target weldment and the preset interval length, and the corresponding welding sub-regions are obtained, which are denoted as the corresponding welding sub-regions of the target weldment.

[0056] The welding quality of the target welded part is detected by the optical camera device in the inspection robot, and the interference fringe image, grayscale image and contour image of each welded sub-region of the target welded part are obtained.

[0057] The interference fringe analysis module is used to analyze the interference fringe images of each welding sub-region of the target weldment, obtain the welding quality reference coefficients corresponding to each welding sub-region, and evaluate the preliminary welding quality assessment results of the target weldment.

[0058] Furthermore, the specific execution method of the interference fringe analysis module is as follows:

[0059] Information is extracted from the interference fringe images of each welding sub-region of the target weldment to obtain the interference fringe information of each welding sub-region of the target weldment. The interference fringe information includes the sharpness of each interference fringe, the first and second spacings of each adjacent interference fringe, and the fringe width and crack number of each interference fringe.

[0060] Temperature detection equipment is used to detect the temperature of the target weldment and obtain the temperature of the welding material in each welding sub-region of the target weldment.

[0061] The ambient light intensity corresponding to the target weldment is obtained through an optical sensor;

[0062] Extract the standard interference fringe datasets stored in the database. The selection data corresponding to the standard interference fringe datasets include ambient light intensity, welding material temperature, and welding material type. Filter the target weldment based on the ambient light intensity, welding material temperature of each welding sub-region, and welding material type to obtain the standard interference fringe datasets for each welding sub-region of the target weldment. The standard interference fringe datasets include standard interference fringe clarity, standard interference fringe spacing, standard interference fringe width, and standard interference fringe crack number.

[0063] By analyzing the interference fringe information of each welding sub-region of the target weldment and the standard interference fringe dataset of each welding sub-region of the target weldment, the welding quality reference coefficient corresponding to each welding sub-region is obtained.

[0064] The welding quality reference coefficients corresponding to each welding sub-region are compared with the preset standard welding quality reference coefficients. If the welding quality reference coefficients corresponding to each welding sub-region are greater than the preset standard welding quality reference coefficients, it indicates that the preliminary welding quality assessment of the welding sub-region is normal. If the welding quality reference coefficients corresponding to each welding sub-region are less than or equal to the preset standard welding quality reference coefficients, it indicates that the preliminary welding quality assessment of the welding sub-region is abnormal. The welding sub-regions with normal preliminary welding quality assessments and the welding sub-regions with abnormal preliminary welding quality assessments corresponding to the target weldment are selected and recorded as the preliminary welding quality assessment results of the target weldment.

[0065] It should be noted that: when a stripe is adjacent to two other stripes, the distance between it and one of the stripes is recorded as the first distance and the distance between it and the other stripe is recorded as the second distance according to the preset stripe coding order. When an edge stripe is adjacent to only one stripe, the first or second distance of the edge stripe is recorded as zero.

[0066] Specifically, through calculation formulas;

[0067] The correlation coefficient between the spacing and width of each interference fringe in each welded sub-region was calculated. ;

[0068] Through calculation formula The sharpness matching coefficient of each interference fringe in each welded sub-region was calculated. ;

[0069] Through calculation formula The crack coincidence coefficient of each interference fringe in each welded sub-region was calculated. ;

[0070] Through calculation formula The welding quality reference coefficients for each welding sub-region were calculated. , This is represented by the number of each welding sub-region. This is indicated by the number of the welding stripe. This is expressed as the number of weld lines. Represented as a natural constant, This is represented by the number of each weld stripe corresponding to each weld sub-region, where... , , , These represent the standard interference fringe spacing, standard interference fringe width, standard interference fringe sharpness, and standard interference fringe crack number for each welding sub-region of the target weldment, respectively. , , , , These represent the sharpness of each interference fringe in each welded sub-region, the first and second spacings of each adjacent interference fringe, the fringe width of each interference fringe, and the number of cracks in each interference fringe, respectively.

[0071] The comprehensive optimization module is used to analyze the grayscale images of each welding sub-region corresponding to the target welded part, obtain the welding quality evaluation coefficient corresponding to each welding sub-region, and perform comprehensive analysis in combination with the preliminary welding quality assessment results of the target welded part to obtain the comprehensive welding quality assessment result of the target welded part.

[0072] Furthermore, the specific execution method of the comprehensive optimization module is as follows:

[0073] By extracting data from the grayscale images of each welding sub-region of the target welded part, the grayscale value of each pixel in each welding sub-region is obtained. The mean, variance and standard deviation of the grayscale value of each pixel in each welding sub-region are calculated.

[0074] The grayscale values ​​of each pixel in each welding sub-region are statistically analyzed to obtain the grayscale histogram corresponding to each welding sub-region. By extracting data from the grayscale histograms corresponding to each welding sub-region, the number of peaks in the grayscale histograms corresponding to each welding sub-region is obtained.

[0075] A data extraction relationship is established between the comprehensive optimization module and the database. The data dispersion model stored in the database is extracted. The data to be filled in corresponding to the data dispersion model includes the mean, variance, standard deviation and grayscale peaks of the pixel grayscale values. Data analysis is performed on the mean, variance, standard deviation and grayscale peaks of the pixel grayscale values ​​in each welding sub-region and the data dispersion model to obtain the grayscale dispersion of the pixel grayscale values ​​in each welding sub-region. The grayscale dispersion of the pixel grayscale values ​​in each welding sub-region is recorded as the welding quality evaluation coefficient corresponding to each welding sub-region.

[0076] It should be noted that the specific implementation method for obtaining the welding quality evaluation coefficients for each welding sub-region based on the data dispersion model analysis is as follows;

[0077] The mean grayscale value of pixels in each welding sub-region is statistically calculated to obtain the sum of the mean grayscale values ​​of pixels in each welding sub-region. Based on the number of welding sub-regions, the average value corresponding to the sum of the mean grayscale values ​​of pixels in each welding sub-region is calculated and denoted as the comprehensive mean grayscale value of pixels in the welding sub-region. The difference between the mean grayscale value of pixels in each welding sub-region and the comprehensive mean grayscale value of pixels in the welding sub-region is calculated to obtain the mean deviation of the grayscale values ​​of pixels in each welding sub-region. For example, if the mean grayscale value of pixels in a certain welding sub-region is A1, and the comprehensive mean grayscale value of pixels in the welding sub-region is... Then, the mean deviation of the grayscale values ​​of the pixels in the welding sub-region = |A1- │;

[0078] The mean deviation, variance, standard deviation, and number of grayscale peaks of pixels in the welding sub-region are respectively labeled as follows: , , P;

[0079] The welding quality evaluation coefficient corresponding to the welding sub-region is:

[0080] ,in It represents the maximum variance across all sub-regions. Represented as the maximum number of peaks across all sub-regions. This represents the maximum deviation of the mean across all sub-regions. This represents the maximum standard deviation across all sub-regions. , , , These represent the influence weights of the variance of pixel grayscale values, the number of peaks in the grayscale histogram, the mean deviation, and the standard deviation of the grayscale values ​​in the welding sub-region on the welding quality evaluation coefficient corresponding to the welding sub-region, respectively. =0.4、 =0.3、 =0.2、 =0.1.

[0081] Extract the standard contour model corresponding to the target welded part stored in the database. Divide the standard contour model corresponding to the target welded part according to each welding sub-region to obtain the standard contour model corresponding to each welding sub-region. Record the standard contour model corresponding to each welding sub-region as the initial contour model corresponding to each welding sub-region.

[0082] The contour images of each welding sub-region are scanned from all directions by a contour scanner to obtain the three-dimensional images corresponding to each welding sub-region, and a post-welding contour model corresponding to each welding sub-region is established.

[0083] Compare the initial contour model corresponding to each welding sub-region with the post-weld contour model corresponding to each welding sub-region to obtain the difference contour model before and after welding for each welding sub-region, obtain the actual volume corresponding to the difference contour model, and statistically obtain the actual volume corresponding to the difference contour model for each welding sub-region.

[0084] Extract the welding shrinkage rates corresponding to various welding materials stored in the database, and obtain the welding shrinkage rates of the welding materials corresponding to the target welding parts by filtering the types of welding materials corresponding to the target welding parts;

[0085] The volume of welding material used is obtained by measuring the amount of welding material used.

[0086] Model analysis was conducted by taking the welding shrinkage rate of the welding material corresponding to the target welded part, the volume of the welding material used for the target welded part, and the actual volume corresponding to the difference contour model of each welding sub-region to obtain the welding material usage compliance coefficient for each welding sub-region.

[0087] By performing data analysis on the welding usage compliance coefficients of the welding materials corresponding to each welding sub-region, the welding quality evaluation coefficients corresponding to each welding sub-region, and the welding quality reference coefficients corresponding to each welding sub-region, the effective welding quality evaluation coefficients corresponding to each welding sub-region are obtained. Among them, the effective welding quality evaluation coefficients corresponding to each welding sub-region are compared with the preset threshold of the effective welding quality evaluation coefficient. If the effective welding quality evaluation coefficients corresponding to each welding sub-region are all greater than or equal to the preset threshold of the effective welding quality evaluation coefficient, it indicates that the welding quality of the target welded part is qualified, and the qualified welding quality of the target welded part is recorded as the comprehensive welding quality evaluation result corresponding to the target welded part. If there is a welding sub-region where the effective welding quality evaluation coefficient is less than the preset threshold of the effective welding quality evaluation coefficient, it indicates that the welding quality of the target welded part is unqualified. The welding sub-region with an effective welding quality evaluation coefficient less than the preset threshold is recorded as each abnormal sub-region, and each abnormal sub-region corresponding to the target welded part is statistically obtained. Each abnormal sub-region corresponding to the target welded part is recorded as the comprehensive welding quality evaluation result corresponding to the target welded part.

[0088] Specifically, through the calculation formula , the actual volume after shrinkage of the weld of the target welded part is calculated , and the theoretical allocated volume of the welding material for each welding sub-region corresponding to the target welded part is obtained through calculation , through the calculation formula , the calculation result is

[0089] The welding usage compliance coefficient of the welding material for each welding sub-region corresponding to the target welded part , where represents the used volume corresponding to the welding material, represents the shrinkage rate of the weld of the welding material corresponding to the target welded part, represents the number of welding sub-regions corresponding to the target welded part; represents the actual volume corresponding to the difference contour model of each welding sub-region;

[0090] Through the calculation formula , the effective welding quality evaluation coefficient corresponding to each welding sub-region is calculated , where represents the welding usage compliance coefficient of the welding material corresponding to each welding sub-region;

[0091] Please refer to Figure 2 , to achieve the above object, the present invention also provides the following technical solution: A robot vision flexible detection method, including the following steps:

[0092] Obtain the task data corresponding to the target welded part;

[0093] By monitoring the image of the welded area of ​​the target welded part, interference fringe images, grayscale images and contour images of each welded sub-region of the target welded part are obtained;

[0094] The interference fringe images of each welding sub-region of the target welded part are analyzed to obtain the welding quality reference coefficients corresponding to each welding sub-region, and the preliminary evaluation results of the welding quality of the target welded part are obtained.

[0095] The grayscale and contour images of each welding sub-region corresponding to the target welded part are analyzed to obtain the welding quality evaluation coefficient and welding material compliance coefficient corresponding to each welding sub-region. Combined with the preliminary welding quality assessment results of the target welded part, a comprehensive analysis is performed to obtain the comprehensive welding quality assessment result of the target welded part.

[0096] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A robot vision flexible inspection system, characterized in that: include: Data acquisition module: used to acquire task data corresponding to the target weldment; Image detection module: used to monitor the image of the welded area of ​​the target welded part, and obtain the interference fringe image, grayscale image and contour image of each welded sub-region of the target welded part; Interference fringe analysis module: used to analyze the interference fringe images of each welding sub-region of the target weldment, obtain the welding quality reference coefficients corresponding to each welding sub-region, and evaluate the preliminary welding quality assessment results of the target weldment; The specific execution method of the interference fringe analysis module is as follows: Information is extracted from the interference fringe images of each welding sub-region of the target weldment to obtain the interference fringe information of each welding sub-region of the target weldment. The interference fringe information includes the sharpness of each interference fringe, the first and second spacings of each adjacent interference fringe, and the fringe width and crack number of each interference fringe. Temperature detection equipment is used to detect the temperature of the target weldment to obtain the temperature of the welding material in each welding sub-region of the target weldment; The ambient light intensity corresponding to the target weldment is obtained through an optical sensor; Extract the standard interference fringe datasets stored in the database. The selection data corresponding to the standard interference fringe datasets include ambient light intensity, welding material temperature, and welding material type. Filter the target weldment based on the ambient light intensity, welding material temperature of each welding sub-region, and welding material type to obtain the standard interference fringe datasets for each welding sub-region of the target weldment. The standard interference fringe datasets include standard interference fringe clarity, standard interference fringe spacing, standard interference fringe width, and standard interference fringe crack number. By analyzing the interference fringe information of each welding sub-region of the target weldment and the standard interference fringe dataset of each welding sub-region of the target weldment, the welding quality reference coefficient corresponding to each welding sub-region is obtained. The welding quality reference coefficients corresponding to each welding sub-region are compared with the preset standard welding quality reference coefficients. If the welding quality reference coefficients corresponding to each welding sub-region are greater than the preset standard welding quality reference coefficients, it indicates that the preliminary welding quality assessment of the welding sub-region is normal. If the welding quality reference coefficients corresponding to each welding sub-region are less than or equal to the preset standard welding quality reference coefficients, it indicates that the preliminary welding quality assessment of the welding sub-region is abnormal. The welding sub-regions with normal preliminary welding quality assessments and the welding sub-regions with abnormal preliminary welding quality assessments corresponding to the target weldment are selected and recorded as the preliminary welding quality assessment results of the target weldment. The comprehensive optimization module is used to analyze the grayscale and contour images of each welding sub-region corresponding to the target welded part, obtain the welding quality evaluation coefficient and welding material compliance coefficient corresponding to each welding sub-region, and perform comprehensive analysis in combination with the preliminary welding quality assessment results of the target welded part to obtain the comprehensive welding quality assessment results of the target welded part. The specific execution method of the comprehensive optimization module is as follows: By extracting data from the grayscale images of each welding sub-region of the target welded part, the grayscale value of each pixel in each welding sub-region is obtained. The mean, variance and standard deviation of the grayscale value of each pixel in each welding sub-region are calculated. The grayscale values ​​of each pixel in each welding sub-region are statistically analyzed to obtain the grayscale histogram corresponding to each welding sub-region. By extracting data from the grayscale histograms corresponding to each welding sub-region, the number of peaks in the grayscale histograms corresponding to each welding sub-region is obtained. A data extraction relationship is established between the comprehensive optimization module and the database. The data dispersion model stored in the database is extracted. The data to be filled in corresponding to the data dispersion model includes the mean, variance, standard deviation and grayscale peaks of the pixel grayscale values. Data analysis is performed on the mean, variance, standard deviation and grayscale peaks of the pixel grayscale values ​​in each welding sub-region and the data dispersion model to obtain the grayscale dispersion of the pixel grayscale values ​​in each welding sub-region. The grayscale dispersion of the pixel grayscale values ​​in each welding sub-region is recorded as the welding quality evaluation coefficient corresponding to each welding sub-region. Extract the standard contour model corresponding to the target welded part stored in the database. Divide the standard contour model corresponding to the target welded part according to each welding sub-region to obtain the standard contour model corresponding to each welding sub-region. Record the standard contour model corresponding to each welding sub-region as the initial contour model corresponding to each welding sub-region. The contour images of each welding sub-region are scanned from all directions by a contour scanner to obtain the three-dimensional images corresponding to each welding sub-region, and a post-welding contour model corresponding to each welding sub-region is established. Compare the initial contour model corresponding to each welding sub-region with the post-weld contour model corresponding to each welding sub-region to obtain the difference contour model before and after welding for each welding sub-region, obtain the actual volume corresponding to the difference contour model, and statistically obtain the actual volume corresponding to the difference contour model for each welding sub-region. Extract the welding shrinkage rates corresponding to various welding materials stored in the database, and obtain the welding shrinkage rates of the welding materials corresponding to the target welding parts by filtering the types of welding materials corresponding to the target welding parts; The volume of welding material used is obtained by measuring the amount of welding material used. Model analysis was conducted by taking the welding shrinkage rate of the welding material corresponding to the target welded part, the volume of the welding material used for the target welded part, and the actual volume corresponding to the difference contour model of each welding sub-region to obtain the welding material usage compliance coefficient for each welding sub-region. By performing data analysis on the welding usage compliance coefficients of the welding materials corresponding to each welding sub-region, the welding quality evaluation coefficients corresponding to each welding sub-region, and the welding quality reference coefficients corresponding to each welding sub-region, the effective welding quality evaluation coefficients corresponding to each welding sub-region are obtained. Among them, the effective welding quality evaluation coefficients corresponding to each welding sub-region are compared with the preset threshold of the effective welding quality evaluation coefficient. If the effective welding quality evaluation coefficients corresponding to each welding sub-region are all greater than or equal to the preset threshold of the effective welding quality evaluation coefficient, it indicates that the welding quality of the target welded part is qualified, and the qualified welding quality of the target welded part is recorded as the comprehensive welding quality evaluation result corresponding to the target welded part. If there is a welding sub-region where the effective welding quality evaluation coefficient is less than the preset threshold of the effective welding quality evaluation coefficient, it indicates that the welding quality of the target welded part is unqualified. The welding sub-region with an effective welding quality evaluation coefficient less than the preset threshold is recorded as each abnormal sub-region, and each abnormal sub-region corresponding to the target welded part is statistically obtained. Each abnormal sub-region corresponding to the target welded part is recorded as the comprehensive welding quality evaluation result corresponding to the target welded part.

2. The robot vision flexible inspection system according to claim 1, characterized in that: The specific execution method of the data acquisition module is as follows: Obtain the task data and welding data of the target welded part through the welding work log; The task data includes the required welding length of the target welded part; The welding data refers to the type and usage amount of the welding material corresponding to the target welded part.

3. The robot vision flexible inspection system according to claim 2, characterized in that: The specific execution method of the image detection module is as follows: Divide the welding area of the target welded part according to the welding length of the target welded part and the preset interval length to obtain each welding sub-region corresponding to the welding area of the target welded part, which is recorded as each welding sub-region corresponding to the target welded part; ​ 4. A robot vision flexibility detection method, applied to the robot vision flexibility detection system according to any one of claims 1-3, characterized in that: ​ ​ ​ ​ ​

Images