Welding quality comprehensive evaluation method and system

Abstract

The invention relates to the field of image processing, in particular to a welding quality comprehensive evaluation method and system, and the method comprises the steps: obtaining an ultrasonic image of a welding joint, carrying out the contour detection, and extracting a defect region of the welding joint; gradient features and gray features of pixel points in the defect area are calculated, feature vectors of target pixel points in the defect area are constructed, and the feature vectors serve as the expression degree of the defect area; edge information of the defect area is obtained so as to determine defect types, the defect types comprise cracks and bubbles, the shape and size of a window are set according to the defect types, bilateral filtering processing is carried out, defect information is extracted, and comprehensive evaluation is carried out on the welding quality. According to the method, the appropriate window shape and size are selected according to the defect type, the detail information of the defect is effectively reserved, the definition of the ultrasonic image is remarkably enhanced through the self-adaptive filtering strategy, and therefore the welding quality is better evaluated.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and more specifically, to a welding quality comprehensive evaluation method and system. Background Art

[0002] Welding technology is a processing technology that connects metal or non-metal materials into one. It is used in various fields such as automobile manufacturing, aerospace, large industrial plants, high-rise buildings, bridges, etc. With the development of modern construction and manufacturing industries, the demand for welding technology continues to increase, and the requirements for welding quality are also increasing. The quality of welding is directly related to the service life and safety of the entire workpiece. By developing a comprehensive evaluation method and system for welding quality, a comprehensive and comprehensive quantitative evaluation of welding quality can be achieved, thereby more accurately evaluating the welding quality level, helping to promptly discover welding quality problems and take corresponding measures to improve them, thereby improving welding quality.

[0003] The existing Chinese patent application document with publication number CN112756768A discloses a welding quality evaluation method and system based on ultrasonic image feature fusion, which relates to the field of welding quality detection technology. It can obtain three-dimensional information of internal defects of welding joints through B-scan and C-scan images detected by ultrasonic detection, and fuse multiple defect feature quantities through information entropy to achieve comprehensive characterization of welding quality, and the evaluation information is comprehensive and accurate; the method includes: S1, ultrasonic detection of backfill stir friction spot welding joints, and collection of B-scan images and C-scan images of welding areas; S2, analyzing and processing the obtained B-scan images and C-scan images, extracting effective information in the images and defect features inside the welds, and forming a three-dimensional information characterization of welding quality; S3, using information entropy to assign different weight coefficients to the feature quantities in the three-dimensional information characterization, and then fusion into a comprehensive indicator as an evaluation result. The technical solution provided by the present invention is applicable to the process of welding quality evaluation.

[0004] The application document obtains three-dimensional information of internal defects of welded joints through B-scan and C-scan images of ultrasonic testing, and uses information entropy to weightedly fuse multiple defect feature quantities, thereby achieving a comprehensive and accurate comprehensive characterization of welding quality, significantly improving the accuracy and reliability of welding quality evaluation. However, due to the propagation characteristics of sound waves in materials or multiple scattering on interfaces, the collected ultrasonic images may be unclear, affecting the extraction of defect information of welded joints, and further affecting the evaluation of welding quality. Summary of the invention

[0005] In order to solve the problem that the acquired ultrasonic image may be unclear due to the propagation characteristics of sound waves in the material or multiple scattering on the interface, thereby affecting the evaluation of welding quality, the present invention provides solutions in the following aspects.

[0006] In a first aspect, a comprehensive evaluation method for welding quality includes: obtaining an ultrasonic image of a welding joint and performing contour detection to extract a defective area of ​​the welding joint; calculating the gradient features and grayscale features of pixels in the defective area, and constructing a feature vector of a target pixel in the defective area, and using the feature vector as the degree of expression of the defective area; obtaining edge information of the defective area to determine the defect type, wherein the defect type includes: cracks and bubbles, setting a window shape and size according to the defect type, performing bilateral filtering and extracting defect information, and comprehensively evaluating the welding quality; wherein the window shape includes: in response to the defect type being a crack, the window of the defective area is preset to a rectangle, and in response to the defect type being a bubble, the window of the defective area is preset to a square; the rectangular window size satisfies the following polynomial: ; ; In the formula, Indicates The length of the pixel window, represents the length of the defect area, represents the width of the defect area, Indicates the crack area The degree to which each pixel represents a defect. Indicates the crack area The width of the window in pixels. Indicates The exponential function with base , represents a round-up function; the square window width satisfies the following relationship: ; In the formula, Indicates The width of the window is pixels. Indicates the crack area The degree to which each pixel represents a defect. Indicates The exponential function with base , Represents the ceiling function.

[0007] The effect is that by combining contour detection, gradient features and grayscale features, the characteristics of the defect area can be more comprehensively described. In particular, for the two common defect types of cracks and bubbles, by analyzing their different geometric and texture features, the defect types can be more accurately distinguished and identified. By calculating the gradient features and grayscale features of the pixels in the defect area, and adapting the window shape and size according to the defect type, the defect information can be extracted more efficiently, making the filtering process more suitable for the actual shape of the defect, thereby reducing unnecessary calculations while retaining the defect details and improving the efficiency of defect information extraction.

[0008] Preferably, the gradient characteristics include: Taking any pixel in the defect area as the target pixel, the sum of square differences between the gradient magnitude and gradient direction of the target pixel and each pixel in the eight neighborhoods is calculated respectively, and the mean of the square root of the sum of square differences is taken as the gradient feature of the target pixel.

[0009] The effect is that by calculating the difference between the gradient size and gradient direction of the target pixel and each pixel in its eight neighborhoods, the edge and structural changes in the defect area can be effectively highlighted. Since cracks usually show significant changes in gradient direction and size, the defect features become more significant, thereby improving the accuracy of defect detection. This method is more sensitive to defects of complex shapes (such as cracks) because cracks usually show significant local changes in gradient size and direction.

[0010] Preferably, the gradient feature further includes: Taking any pixel in the defect area as the target pixel, the difference between the gradient magnitude of the target pixel and the average value of the gradient magnitudes of all pixels in the neighborhood is calculated to obtain the difference in gradient magnitudes, and the difference between the gradient direction of the target pixel and the average value of the gradient directions of all pixels in the neighborhood is calculated to obtain the difference in gradient directions; The square root of the sum of the squares of the difference in gradient magnitude and the difference in gradient direction is taken as the gradient feature of the target pixel.

[0011] The effect is that by calculating the difference between the target pixel and the neighborhood average, the overall characteristics of the target pixel relative to its neighborhood are directly reflected. The calculation process is simple and more efficient, suitable for large-scale image processing and real-time detection. This method is more suitable for bubble defects with regular shapes, because the gradient changes of regular defects are relatively uniform, and the difference with the neighborhood average can better reflect its characteristics.

[0012] Preferably, the grayscale features include: Taking any pixel in the defect area as the target pixel, the difference between the grayscale value of the target pixel and the grayscale mean in the defect area is calculated, and the difference is divided by the absolute value of the difference between the maximum and minimum grayscale values ​​in the defect area as the grayscale feature of the target pixel.

[0013] The effect is that: through normalization processing, the difference between the grayscale value of the target pixel and the grayscale mean of the defect area is standardized, so that it is not affected by the specific grayscale range, the robustness to noise is enhanced, and the significant features in the defect area can be effectively highlighted.

[0014] Preferably, the grayscale feature further includes: Taking any pixel in the defect area as the target pixel, the difference between the grayscale value of the target pixel and the grayscale variance in the defect area is calculated, and the difference is divided by the absolute value of the difference between the maximum and minimum grayscale values ​​in the defect area as the grayscale feature of the target pixel.

[0015] Preferably, the step of obtaining edge information of the defect area for determining the defect type includes: Obtain the length and width of the defect area, where the length is the maximum value of the distance between the two ends of the pixel points in the horizontal direction of the defect area, and the width is the maximum value of the distance between the two ends of the pixel points in the vertical direction of the defect area. According to the difference between the length and width of the pixel points, obtain the appearance of the defect area; The absolute difference between the length and width of the defect area is compared with the preset parameters. If the absolute difference is greater than or equal to the preset parameters, the length and width difference of the defect area is large and it is judged to be a crack. Otherwise, the length and width difference is small and it is judged to be a bubble.

[0016] The effect is that the judgment method based on length and width difference can effectively distinguish cracks and bubbles, reducing misjudgment caused by similar shapes. Cracks and bubbles have different effects in welding quality assessment, which is conducive to better assessment of welding quality.

[0017] Preferably, the step of obtaining edge information of the defect area for determining the defect type further includes: Based on contour detection, the boundary of the defective area is extracted, the boundary curvature of each pixel on the boundary is calculated, the defect type is judged according to the statistical characteristics of the boundary curvature, the average value and standard deviation of the boundary curvature are calculated respectively, and the average value and standard deviation of the boundary curvature are compared with the corresponding preset parameters respectively. In response to the average value and standard deviation being greater than or equal to the corresponding preset parameters, the curvature characteristics of the defective area are consistent with the crack and it is judged to be a crack. Otherwise, it is judged to be a bubble.

[0018] In a second aspect, a welding quality comprehensive evaluation system includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned welding quality comprehensive evaluation method is implemented.

[0019] The present invention has the following effects: 1. This method selects the appropriate window shape and size according to the defect type, which can effectively retain the detailed information of the defect while removing noise and blur. The adaptive filtering strategy significantly enhances the clarity of the ultrasonic image, making the defect information easier to extract and analyze, thereby improving the efficiency and quality of weld joint defect detection.

[0020] 2. The present invention combines gradient features and grayscale features to analyze the degree of defect expression of pixels in the defect area. For the two common defect types of cracks and bubbles, by analyzing their different geometric and texture features, the defect types can be more accurately distinguished and identified. This multi-feature fusion method significantly improves the accuracy and reliability of defect detection, thereby providing a more accurate basis for welding quality assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 It is a method flow chart of steps S1 to S3 in a method for comprehensive evaluation of welding quality in an embodiment of the present invention.

[0022] Figure 2 It is a structural block diagram of a comprehensive welding quality evaluation system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0024] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0025] Reference Figure 1 A comprehensive evaluation method for welding quality includes steps S1 to S3, which are as follows: Specific implementation scenario: The three-dimensional information of the internal defects of the welded joint is obtained through the B-scan and C-scan images of ultrasonic testing. When evaluating the welding quality, the clarity of the defect area is very important, which is related to the subsequent evaluation of the welding quality. When using ultrasound to collect images of the welded joint area, the propagation characteristics of the sound waves in the material or multiple scattering on the interface may cause the collected ultrasonic image to be unclear, affecting the extraction of defect information of the welded joint, and further affecting the evaluation of its welding quality. Therefore, it is necessary to filter the ultrasonically collected welded joint area image to improve the performance of the defect area, which is helpful for the subsequent welding quality evaluation.

[0026] S1: Obtain an ultrasonic image of the weld joint and perform contour detection to extract the defect area of ​​the weld joint.

[0027] In this embodiment, the contour detection adopts Canny edge detection, which is a well-known technology in the art and will not be described in detail.

[0028] It should be noted that in the defect area, due to the difference in physical properties between the defect and the surrounding normal tissue, the image brightness often changes significantly, and this change is manifested as an increase in the gradient value on the gradient map. In other words, the gradient feature can capture the edge information of the defect, which is of great significance for identifying the outline and shape of the defect. At the same time, there is often a difference in the grayscale value between the defect area and the surrounding normal tissue. The grayscale feature can be used as a supplement to the gradient feature to jointly analyze the degree of defect expression of the pixel point.

[0029] S2: Calculate the gradient features and grayscale features of the pixels in the defect area, and construct a feature vector of the target pixel in the defect area, and use the feature vector as the representation degree of the defect area.

[0030] Embodiment 1: Obtaining the degree of representation of defective areas includes: Taking any pixel in the defect area as the target pixel, the sum of square differences between the gradient magnitude and gradient direction of the target pixel and each pixel in the eight neighborhoods is calculated respectively, and the mean of the square root of the sum of square differences is taken as the gradient feature of the target pixel. Specifically, the gradient feature of the target pixel satisfies the following relationship: ; In the formula, Indicates The gradient features of pixels, Indicates the defect area The gradient size of each pixel, Indicates the defect area In the neighborhood of pixels The gradient size of each pixel, Indicates the defect area The gradient direction of each pixel is Indicates the defect area In the neighborhood of pixels The gradient direction of each pixel.

[0031] That is to say, Indicates The pixel and the neighborhood The difference in the intensity of brightness changes of the pixels. The greater the difference in gradient size, the greater the difference in the intensity of brightness changes; Indicates The pixel and the neighborhood The difference in the direction of brightness change of each pixel. The greater the difference in the gradient direction, the greater the difference in the direction of brightness change.

[0032] Taking any pixel in the defect area as the target pixel, calculating the difference between the grayscale value of the target pixel and the grayscale mean value in the defect area, and dividing the difference by the absolute value of the difference between the maximum and minimum grayscale values ​​in the defect area as the grayscale feature of the target pixel; Specifically, the grayscale feature of the target pixel satisfies the following relationship: ; In the formula, Indicates The grayscale features of pixels, Indicates the defect area The gray value of a pixel, represents the grayscale mean in the defect area, Indicates the maximum value of the gray value in the defect area, Indicates the minimum grayscale value in the defect area.

[0033] That is to say, The role of normalization is to normalize the difference Normalized to This makes the grayscale features of pixels in different defect areas comparable and is not affected by the specific grayscale value range. The larger the The greater the difference between the grayscale value of each pixel and the average grayscale value in the defect area, the more likely it is the edge or internal area of ​​the defect.

[0034] The gradient features and grayscale features are used to construct the feature vector of the target pixel in the defect area as the representation degree of the defect area.

[0035] Specifically, the degree of defect expression satisfies the following relationship: ; In the formula, Indicates The degree of pixel defect performance, Indicates the defect area The feature vector of pixels, Indicates the defect area The gradient features of pixels, Indicates the defect area The grayscale features of pixels, Represents the magnitude of a vector.

[0036] That is to say, the defect area The larger the gradient feature of a pixel, the more dramatic the brightness change of the pixel. It is likely to be the edge or internal area of ​​the defect, and the more obvious the defect is. The larger the grayscale feature of a pixel, the more significantly different the brightness level of the pixel is from the defect area, and the greater the impact of the defect. That is, the larger the modulus of the feature vector composed of the grayscale feature and the gradient feature of the pixel, the greater the degree to which the pixel expresses the defect.

[0037] In addition, the second embodiment also includes: Taking any pixel in the defect area as the target pixel, the difference between the gradient magnitude of the target pixel and the average value of the gradient magnitudes of all pixels in the neighborhood is calculated to obtain the difference in gradient magnitudes, and the difference between the gradient direction of the target pixel and the average value of the gradient directions of all pixels in the neighborhood is calculated to obtain the difference in gradient directions; The square root of the sum of the squares of the difference in gradient magnitude and the difference in gradient direction is taken as the gradient feature of the target pixel.

[0038] Specifically, the gradient feature of the target pixel satisfies the following relationship: ; In the formula, Indicates The gradient features of pixels, Indicates the defect area The gradient size of each pixel, Indicates the defect area In the neighborhood of pixels The gradient size of each pixel, Indicates the defect area The gradient direction of each pixel is Indicates the defect area In the neighborhood of pixels The gradient direction of each pixel.

[0039] Taking any pixel in the defect area as the target pixel, the difference between the grayscale value of the target pixel and the grayscale variance of the defect area is calculated, and the difference is divided by the absolute value of the difference between the maximum and minimum grayscale values ​​in the defect area as the grayscale feature of the target pixel.

[0040] Specifically, the grayscale feature of the target pixel satisfies the following relationship: ; In the formula, Indicates The grayscale features of pixels, Indicates the defect area The gray value of a pixel, represents the grayscale variance in the defect area, Indicates the maximum value of the gray value in the defect area, Indicates the minimum grayscale value in the defect area.

[0041] Among them, the difference between the second embodiment and the first embodiment is: the calculation method of gradient features and grayscale features. The gradient feature calculation in the first embodiment is more sensitive to defects of complex shapes, such as cracks, because cracks usually show significant local changes in gradient size and direction, while the gradient feature calculation in the second embodiment is more suitable for bubble defects of regular shapes, because the gradient changes of regular defects are relatively uniform, and their characteristics can be better reflected by the difference with the neighborhood average. Although complex defects such as cracks can also be detected, they may not be as sensitive as the first embodiment, and can be selected according to specific circumstances to achieve better detection effects.

[0042] To further explain, common defects in the welding process are cracks and bubbles. Different defects have different manifestations, and the shapes of the required windows are also different. Information on the defect shape helps determine the appropriate filter window shape and size. Cracks usually appear as thin and long lines, while bubbles may appear as circles or ovals. For crack defects, using a rectangular filter window can better cover the extension direction of the crack, thereby more accurately capturing crack information; and for bubble defects, using a square filter window may be more appropriate, because bubbles usually appear as more regular circles or ovals.

[0043] S3: Obtain edge information of the defect area to determine the defect type, where the defect types include: cracks and bubbles, set the window shape and size according to the defect type, perform bilateral filtering and extract defect information, and conduct a comprehensive evaluation of the welding quality.

[0044] In this embodiment, bilateral filtering is a technology well known to those skilled in the art and will not be described in detail. The ultrasonic image of the welded joint is filtered and a filtering model is trained. The trained model is installed in the welding quality assessment system to improve the assessment effect.

[0045] The specific steps of extracting defect information after bilateral filtering, fusing multiple defect feature quantities, and comprehensively evaluating welding quality refer to the invention of a welding quality evaluation method and system based on ultrasonic image feature fusion with announcement number CN112756768B, which has detailed steps and will not be described in detail. The present invention mainly filters ultrasonic images for different defect types and adaptively adjusts the window size. The specific steps are as follows: Obtain the length and width of the defect area, where the length is the maximum value of the distance between the two ends of the pixel points in the horizontal direction of the defect area, and the width is the maximum value of the distance between the two ends of the pixel points in the vertical direction of the defect area. According to the difference between the length and width of the pixel points, obtain the appearance of the defect area; The absolute difference between the length and width of the defect area is compared with the preset parameters. If the absolute difference is greater than or equal to the preset parameters, the length and width difference of the defect area is large and it is judged to be a crack. Otherwise, the length and width difference is small and it is judged to be a bubble.

[0046] Specifically, the defect type satisfies the following relationship: ; In the formula, Indicates the defect type, represents the length of the defect area, represents the width of the defect area, Indicates preset parameters.

[0047] That is to say, in this embodiment When the difference between the length and width of the defect area exceeds the preset parameters, it is considered that the difference between the length and width of the defect area is large, and the possibility that the shape of the defect area is a rectangle is greater, and the possibility that the defect is a crack is greater; conversely, when the difference between the length and width of the defect area does not exceed the preset parameters, it is considered that the difference between the length and width of the defect area is small, and the possibility that the shape of the defect area is a square is greater, and the possibility that the defect is a bubble is greater.

[0048] In addition, another embodiment further includes: Based on contour detection, the boundary of the defective area is extracted, the boundary curvature of each pixel on the boundary is calculated, the defect type is judged according to the statistical characteristics of the boundary curvature, the average value and standard deviation of the boundary curvature are calculated respectively, and the average value and standard deviation of the boundary curvature are compared with the corresponding preset parameters respectively. In response to the average value and standard deviation being greater than or equal to the corresponding preset parameters, the curvature characteristics of the defective area are consistent with the crack and it is judged to be a crack. Otherwise, it is judged to be a bubble.

[0049] Exemplarily, the preset parameters are 0.1 and 0.05 respectively; the average curvature reflects the overall curvature of the boundary. A higher average curvature indicates that the boundary is more complex, such as a crack; the curvature standard deviation reflects the degree of change in the boundary curvature. A higher standard deviation indicates that the boundary curvature changes greatly, such as a crack.

[0050] Characteristics of cracks: Cracks are usually formed by stress concentration or defect expansion inside the material. Their boundary shapes are relatively complex and may include bends in multiple directions, sharp turning points, and irregular shapes. As a result, the curvature varies greatly at different locations. The tip or turning point of the crack usually has a higher curvature, while the straight part of the crack has a lower curvature.

[0051] Characteristics of bubbles: Bubbles are usually formed by gas or liquid encapsulation inside the material. Their boundary shape is relatively regular, usually close to a circle or an ellipse, and the overall curvature is low.

[0052] The window shape includes: in response to the defect type being a crack, the window of the defect area is preset to be a rectangle; in response to the defect type being a bubble, the window of the defect area is preset to be a square; The rectangular window size satisfies the following relationship: ; ; In the formula, Indicates The length of the pixel window, represents the length of the defect area, represents the width of the defect area, Indicates the crack area The degree to which each pixel represents a defect. Indicates the crack area The width of the window in pixels. Indicates The exponential function with base , represents the ceiling function; That is to say, the first The greater the degree to which each pixel represents the defect, the smaller the window required. For the normalization process, is the difference between the length and width of the defect area. The larger the difference, the more obvious the rectangular feature of the window, and the greater the impact on the setting of the window length of the pixel point. When the length of the defect area is greater than the width, the window of the pixel point must also meet this feature to ensure that the remaining pixels in the pixel point window also belong to the defect area. Therefore, the length of the pixel point window must also be greater than the width. At this time, the width of the pixel point window is half of the length. Otherwise, the width of the pixel point window is twice the length.

[0053] The square window width satisfies the following relationship: ; In the formula, Indicates The width of the window in pixels. The first The degree to which each pixel represents a defect. Indicates The exponential function with base , Represents the ceiling function.

[0054] That is to say, within the bubble area The more a pixel represents a defect, the smaller the window required.

[0055] The present invention also provides a comprehensive evaluation system for welding quality. Figure 2 As shown, the system includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a method for comprehensive evaluation of welding quality according to the first aspect of the present invention is implemented.

[0056] The system also includes other components familiar to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and will not be described in detail here.

[0057] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of a device or accessible or connectable to a device. Any application or module described in the present invention may be implemented by computer-readable / executable instructions stored or otherwise maintained by such a computer-readable medium.

[0058] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0059] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A comprehensive evaluation method for welding quality, characterized in that: include: Obtain ultrasonic images of welded joints and perform contour detection to extract defective areas of welded joints; Calculate the gradient features and grayscale features of the pixels in the defect area, and construct the feature vector of the target pixel in the defect area, and use the feature vector as the representation degree of the defect area; Obtain edge information of the defect area to determine the defect type, wherein the defect types include: cracks and bubbles, set the window shape and size according to the defect type, perform bilateral filtering and extract defect information, and conduct a comprehensive evaluation of the welding quality; wherein the window shape includes: in response to the defect type being a crack, the window of the defect area is preset to a rectangle, and in response to the defect type being a bubble, the window of the defect area is preset to a square; the rectangular window size satisfies the following polynomial: ; ; In the formula, Indicates The length of the pixel window, represents the length of the defect area, represents the width of the defect area, Indicates the crack area The degree to which each pixel represents a defect. Indicates the crack area The width of the window in pixels. Indicates The exponential function with base , represents the ceiling function; The square window width satisfies the following relationship: ; In the formula, Indicates The width of the window is pixels. Indicates the crack area The degree to which each pixel represents a defect. Indicates The exponential function with base , Represents the ceiling function.

2. A comprehensive evaluation method for welding quality according to claim 1, characterized in that: The gradient features include: Taking any pixel in the defect area as the target pixel, the sum of square differences between the gradient magnitude and gradient direction of the target pixel and each pixel in the eight neighborhoods is calculated respectively, and the mean of the square root of the sum of square differences is taken as the gradient feature of the target pixel.

3. A comprehensive evaluation method for welding quality according to claim 1, characterized in that: The gradient features also include: Taking any pixel in the defect area as the target pixel, the difference between the gradient magnitude of the target pixel and the average value of the gradient magnitudes of all pixels in the neighborhood is calculated to obtain the difference in gradient magnitudes, and the difference between the gradient direction of the target pixel and the average value of the gradient directions of all pixels in the neighborhood is calculated to obtain the difference in gradient directions; The square root of the sum of the squares of the difference in gradient magnitude and the difference in gradient direction is taken as the gradient feature of the target pixel.

4. A comprehensive evaluation method for welding quality according to claim 1, characterized in that: The grayscale features include: Taking any pixel in the defect area as the target pixel, the difference between the grayscale value of the target pixel and the grayscale mean in the defect area is calculated, and the difference is divided by the absolute value of the difference between the maximum and minimum grayscale values ​​in the defect area as the grayscale feature of the target pixel.

5. A comprehensive evaluation method for welding quality according to claim 1, characterized in that: The grayscale features also include: Taking any pixel in the defect area as the target pixel, the difference between the grayscale value of the target pixel and the grayscale variance in the defect area is calculated, and the difference is divided by the absolute value of the difference between the maximum and minimum grayscale values ​​in the defect area as the grayscale feature of the target pixel.

6. A comprehensive evaluation method for welding quality according to claim 1, characterized in that: The obtaining edge information of the defect area to determine the defect type includes: Obtain the length and width of the defect area, where the length is the maximum value of the distance between the two ends of the pixel points in the horizontal direction of the defect area, and the width is the maximum value of the distance between the two ends of the pixel points in the vertical direction of the defect area. According to the difference between the length and width of the pixel points, obtain the appearance of the defect area; The absolute difference between the length and width of the defect area is compared with the preset parameters. If the absolute difference is greater than or equal to the preset parameters, the length and width difference of the defect area is large and it is judged to be a crack. Otherwise, the length and width difference is small and it is judged to be a bubble.

7. A comprehensive evaluation method for welding quality according to claim 1, characterized in that: The step of obtaining edge information of the defect area for determining the defect type further includes: Based on contour detection, the boundary of the defective area is extracted, the boundary curvature of each pixel on the boundary is calculated, the defect type is judged according to the statistical characteristics of the boundary curvature, the average value and standard deviation of the boundary curvature are calculated respectively, and the average value and standard deviation of the boundary curvature are compared with the corresponding preset parameters respectively. In response to the average value and standard deviation being greater than or equal to the corresponding preset parameters, the curvature characteristics of the defective area are consistent with the crack and it is judged to be a crack. Otherwise, it is judged to be a bubble.

8. A welding quality comprehensive evaluation system, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the comprehensive evaluation method for welding quality according to any one of claims 1 to 7 is implemented.

Images