Welding quality visual inspection method and system applied to welding machine

Abstract

The present invention relates to the field of image processing technology, and in particular to a welding quality visual inspection method and system applied to a welding machine. The method performs semantic segmentation processing on a welding image to obtain an object area image and a background area image; performs edge detection on the object area image, and determines the welding area according to the grayscale value change of the area formed by continuous edge points; determines the welding quality factor according to the height distribution influencing factor of the weld corresponding to the welding area, the weld width performance influencing factor and the texture uniformity, and comprehensively judges the welding quality by comprehensively considering the height distribution of the weld, the weld width performance and the texture uniformity of the weld, thereby ensuring the accuracy of the quality judgment; in response to the welding quality factor being greater than the quality threshold, the weld corresponding to the welding area passes the quality inspection, thereby completing the quality inspection. The method comprehensively considers the height distribution of the weld, the weld width performance and the texture uniformity of the weld to judge the welding quality, thereby improving the accuracy of the quality inspection.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a welding quality visual detection method and system applied to a welding machine. Background Art

[0002] With the rapid development of machine vision and image processing technology, visual inspection technology is increasingly used in the field of industrial manufacturing. These technologies not only improve the accuracy and efficiency of detection, but also realize non-contact and non-destructive testing, providing strong technical support for real-time monitoring of welding quality. Welding is an important connection process in industrial manufacturing, and its quality directly affects the safety performance and service life of the product. In industries such as aerospace, automobile manufacturing, and shipbuilding, welding quality is directly related to the overall performance and reliability of the product. Therefore, strict testing and monitoring of welding quality is an important part of ensuring product quality and safety.

[0003] Traditional welding quality inspection methods, such as visual inspection and physical inspection (such as radiographic inspection and ultrasonic inspection), can detect welding defects to a certain extent, but they have limitations such as low inspection efficiency, low accuracy, and susceptibility to human factors, making it difficult to meet the needs of efficient and accurate inspection. Therefore, a welding quality visual inspection method and system for welding machines is urgently needed to solve the above problems. Summary of the invention

[0004] In order to solve the technical problem of inaccurate welding quality detection, the purpose of the present invention is to provide a welding quality visual detection method and system applied to a welding machine. The technical solution adopted is as follows:

[0005] In a first aspect, the present invention provides a welding quality visual inspection method applied to a welding machine, the method comprising:

[0006] Acquire welding images;

[0007] Performing semantic segmentation processing on the welding image to obtain an object area image and a background area image, wherein the object area image includes the object to be measured, and the background area image includes background information of the photographed object;

[0008] Performing edge detection on the object area image, and determining the welding area according to the gray value change of the area formed by continuous edge points, wherein the object area image includes a top view and a side view of the object to be measured, and the welding area includes a top view area and a side view area of ​​a weld on the object to be measured, the top view area of ​​the weld contains weld bead width information and texture information of the weld, and the side view area of ​​the weld contains height information of the weld;

[0009] Determine the welding quality factor according to the height distribution influencing factor of the weld corresponding to the welding area, the weld width performance influencing factor and the texture uniformity;

[0010] In response to the welding quality factor being greater than the quality threshold, the weld corresponding to the welding area passes the quality inspection.

[0011] In some embodiments, the welding quality factor is determined according to the height distribution influencing factor of the weld corresponding to the welding area, the weld width performance influencing factor and the texture uniformity, including:

[0012] Determine the height distribution influencing factor of the weld corresponding to the welding area according to the difference between the height of the weld corresponding to the welding area at various locations and the average height;

[0013] Determine a weld bead width performance influencing factor of the weld corresponding to the welding area according to the variance of the width of each weld bead corresponding to the welding area and the difference in angles between adjacent edge pixels of the welding area;

[0014] Determining a weld shape influencing factor based on the weld width performance influencing factor and the height distribution influencing factor;

[0015] The welding quality factor is determined according to the texture uniformity and shape influence factor of the weld corresponding to the welding area.

[0016] In some embodiments, the height distribution influencing factor of the weld corresponding to the welding area is determined according to the following formula:

[0017]

[0018] In the formula, Indicates the influence factor of the height distribution of the weld, Indicates the side view of the object to be measured. The overall height of the object and weld, Indicates the side view of the object to be measured. The height of the object itself, represents the average height of the weld, represents the position number, and N represents the number of positions.

[0019] In some embodiments, the influencing factor of the weld width of the weld corresponding to the welding area is determined:

[0020]

[0021] In the formula, Indicates the influencing factor of the weld bead width. The top view area of ​​the weld The difference in angles between adjacent edge pixels, It represents the variance of the width of the weld bead corresponding to all edge pixels in the weld top-view area in the horizontal direction. represents the function to find the maximum value, Represents the logarithm of adjacent edge pixels; It is the maximum value of the difference in angles between adjacent edge pixels in the weld top-view area.

[0022] In some embodiments, the difference in angles between adjacent edge pixels of the welding area is determined according to the following formula:

[0023]

[0024] In the formula, The top view area of ​​the weld The difference in angles between adjacent edge pixels, The zth pixel in the weld top view area The angle between the line connecting the edge pixel point and the z-th edge pixel point and the horizontal right direction, It represents the angle between the line between the zth edge pixel point and the z+1th edge pixel point in the weld top-view area and the horizontal right direction, where z represents the number of edge pixels in the weld top-view area.

[0025] In some embodiments, the shape influence factor of the weld is determined according to the following formula:

[0026]

[0027] In the formula, represents the influence factor of weld shape, Indicates the influence factor of the height distribution of the weld, Indicates the influencing factor of the weld bead width.

[0028] In some embodiments, the welding quality factor is determined according to the following formula:

[0029]

[0030] In the formula, represents the welding quality factor, represents the influence factor of weld shape, Indicates the first The pixel value of a pixel, Represents the average value of all pixel values ​​on the feature map corresponding to the weld top view area, Indicates the number of pixels on the feature map corresponding to the weld top view area. Indicates the number of pixels on the feature map corresponding to the weld top view area, Represents the normalization function.

[0031] In some embodiments, a feature map corresponding to a top-view area of ​​a weld is obtained by a local binary method, and a texture uniformity of the weld corresponding to the welding area is determined based on a pixel value of each pixel point in the feature map.

[0032] In some embodiments, edge detection is performed on the object region image, and determining the welding region according to the gray value change of the region formed by continuous edge points includes:

[0033] Performing edge detection on the object area image to obtain an edge detection result;

[0034] In the edge detection result, the grayscale value of the area formed by the continuous edge points is determined to obtain a grayscale value sequence of the area formed by the continuous edge points;

[0035] In response to the fact that the difference between a certain grayscale value in the grayscale value sequence and its adjacent grayscale values ​​is greater than a first grayscale difference threshold, the area formed by the continuous edge points corresponding to the grayscale value is determined to be the welding area, and the number of grayscale values ​​whose differences with its adjacent grayscale values ​​are greater than the first grayscale difference threshold is the number of welding areas, or, the number of pixels in the area formed by the continuous edge points is sorted in order from small to large to obtain a pixel number sorting series of the area formed by the continuous edge points, and in response to the fact that the difference in grayscale values ​​of the area formed by the continuous edge points corresponding to the number of pixels in the pixel number sorting series and the number of adjacent pixels is greater than a second grayscale difference threshold, the area formed by the continuous edge points corresponding to the grayscale value is determined to be the welding area.

[0036] In a second aspect, the present invention provides a welding quality visual inspection system for a welding machine, the system comprising:

[0037] An acquisition module, used for acquiring welding images;

[0038] A segmentation module, used for performing semantic segmentation processing on the welding image to obtain an object area image and a background area image, wherein the object area image includes the object to be measured, and the background area image includes background information of the photographed object;

[0039] A region detection module is used to perform edge detection on the object region image, and determine the welding region according to the gray value change of the region formed by continuous edge points, wherein the object region image includes a top view and a side view of the object to be measured, and the welding region includes a top view region and a side view region of a weld on the object to be measured, the top view region of the weld contains weld bead width information and texture information of the weld, and the side view region of the weld contains height information of the weld;

[0040] A determination module, used to determine a welding quality factor according to a height distribution influencing factor of a weld corresponding to the welding area, a weld width performance influencing factor and a texture uniformity;

[0041] The quality detection module is used for responding that the welding quality factor is greater than the quality threshold and the weld corresponding to the welding area passes the quality detection.

[0042] The present invention has the following beneficial effects:

[0043] The present invention proposes a welding quality visual inspection method and system applied to a welding machine. The method first obtains a welding image to serve as a basis for subsequent welding quality inspection; then, the welding image is semantically segmented to obtain an object area image and a background area image, so as to facilitate the subsequent acquisition of information for analysis directly from the object area image; then, the object area image is edge detected, and the welding area is determined according to the gray value change of the area formed by continuous edge points, so as to accurately lock the welding area to be quality inspected; further, the welding quality factor is determined according to the high and low distribution influencing factor of the weld corresponding to the welding area, the weld width performance influencing factor and the texture uniformity, and the welding quality is judged by comprehensively considering the high and low distribution of the weld, the weld width performance and the texture uniformity of the weld, so as to ensure the accuracy of quality judgment; finally, in response to the welding quality factor being greater than the quality threshold, the weld corresponding to the welding area passes the quality inspection, thereby completing the quality inspection. The method comprehensively considers the high and low distribution of the weld, the weld width performance and the texture uniformity of the weld to judge the welding quality, thereby improving the accuracy of quality inspection. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0045] Figure 1 A method flow chart of a method for visually inspecting welding quality of a welding machine provided by an embodiment of the present invention;

[0046] Figure 2 A top view of an object to be tested in a welding quality visual inspection method for a welding machine provided by an embodiment of the present invention;

[0047] Figure 3 A side view of an object to be tested in a welding quality visual inspection method for a welding machine provided by an embodiment of the present invention;

[0048] Figure 4 A schematic diagram of weld height of a welding quality visual inspection method applied to a welding machine provided by one embodiment of the present invention;

[0049] Figure 5 A schematic diagram of the structure of a welding quality visual inspection system for a welding machine provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0050] 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0052] A specific scheme of a welding quality visual inspection method and system for a welding machine provided by the present invention will be described in detail below in conjunction with the accompanying drawings.

[0053] In the first aspect, the present application provides a method for visually inspecting welding quality applied to a welding machine, see Figure 1 , the method comprising:

[0054] Step 101, acquiring a welding image.

[0055] The welding image is an image of the object to be tested including the weld, and is the basis for subsequent welding quality inspection.

[0056] In some embodiments, a high-resolution industrial camera and a precision optical system are used to capture images of the object to be tested containing a weld. The specific acquisition method is as follows: the optical system is set so that the light source is horizontally irradiated, parallel to the shooting direction of the camera; a high-resolution industrial camera is used to capture top and side views of the object to be tested containing a weld.

[0057] In some embodiments, the welding image can be subjected to histogram equalization processing, and the pixel points of the image can be redistributed to make the histogram of the welding image more uniform, thereby enhancing the overall contrast of the image and making the features of each area on the image more prominent, which is convenient for subsequent segmentation.

[0058] Step 102, performing semantic segmentation processing on the welding image to obtain an object area image and a background area image, wherein the object area image includes the object to be measured, and the background area image includes background information of the photographed object.

[0059] The object area image is an image of the area where the object to be tested including the weld is located, and the background area image is a background image of other areas except the area where the object to be tested including the weld is located. The semantic segmentation processing can segment the object area and the background area in the welding image to obtain the object area image and the background area image, which is convenient for subsequent direct acquisition of information for analysis from the object area image.

[0060] Poor welding quality of an object may have the following defects, including undercut, weld nodules and arc pits. The present invention mainly analyzes the poor quality caused by weld shape problems. Among them, shape problems include uneven weld shape, uneven weld width, too large or too small weld excess height, too wide or too narrow weld width, and uneven transition between weld and parent material. According to the difference between the weld and the normal part in the welding image and the specific characteristics of the weld, the analysis is carried out to accurately lock the welding area to be quality inspected.

[0061] Step 103, edge detection is performed on the object area image, and the welding area is determined according to the gray value change of the area formed by the continuous edge points, wherein, see Figure 2 and Figure 3 The object area image includes a top view and a side view of the object to be measured, and the welding area includes a top view area and a side view area of ​​the weld on the object to be measured. The top view area of ​​the weld contains the weld bead width information and texture information of the weld, and the side view area of ​​the weld contains the height information of the weld.

[0062] For the weld on the object to be tested, there are certain differences between it and other normal areas on the object to be tested. The weld area is welded by welding wire, which makes the weld area show greater irregularity, which is quite different from other normal areas on the object to be tested. The ideal weld area should be flush and the weld width should be uniform, which helps to ensure the safety performance and service life of the product. Therefore, it is necessary to first determine the location of the weld area on the object area image.

[0063] In some embodiments, step 103 specifically includes:

[0064] Step 1031, performing edge detection on the object area image to obtain an edge detection result.

[0065] Edge detection can make the object area image show the contour information of the welding area and the object to be measured itself, which is the basis for subsequent analysis of the difference between the weld and the normal part in the welding image and the specific characteristics of the weld.

[0066] Step 1032: in the edge detection result, determine the grayscale value of the area formed by the continuous edge points, and obtain a grayscale value sequence of the area formed by the continuous edge points.

[0067] Since the surface of the welding area is relatively regular and generally appears silvery white, and the area where the object to be measured is relatively regular and generally appears as the color of the object, the continuous edge points can represent the regional contour information of the welding area and the object to be measured itself, and the grayscale value of the area formed by the continuous edge points can present the overall grayscale of the welding area and the object to be measured itself, so as to subsequently analyze the difference between the grayscale values ​​of the area formed by the continuous edge points and determine the welding area.

[0068] Step 1033, in response to the difference between a certain grayscale value in the grayscale value sequence and its adjacent grayscale values ​​being greater than a first grayscale difference threshold, determining that the area formed by the continuous edge points corresponding to the grayscale value is the welding area, and the number of grayscale values ​​whose differences with its adjacent grayscale values ​​are greater than the first grayscale difference threshold is the number of welding areas, or, sorting the number of pixels in the area formed by the continuous edge points in order from small to large to obtain a pixel number sorting sequence of the area formed by the continuous edge points, in response to the difference in the grayscale value of the area formed by the continuous edge points corresponding to the number of pixels in the pixel number sorting sequence and the number of adjacent pixels being greater than a second grayscale difference threshold, determining that the area formed by the continuous edge points corresponding to the grayscale value is the welding area.

[0069] Step 1033 provides two methods for determining the welding area based on the change in grayscale value of the area formed by continuous edge points.

[0070] In the first method, the difference between a grayscale value in the grayscale value sequence and its adjacent grayscale value is greater than the first grayscale difference threshold, indicating that the area formed by the continuous edge points corresponding to the grayscale value has a large difference with its adjacent grayscale value, and the area formed by the continuous edge points corresponding to the grayscale value is the welding area.

[0071] In the second method, the number of pixels in the area formed by continuous edge points is first sorted to facilitate more accurate locking of the area formed by the continuous edge points corresponding to the number of pixels in the area formed by the continuous edge points. The difference in grayscale values ​​between the number of pixels in the sorted series of pixel numbers and the area formed by the continuous edge points corresponding to the number of its adjacent pixels is greater than the second grayscale difference threshold, indicating that the grayscale value difference between the area formed by the continuous edge points corresponding to the number of pixels and the area formed by the continuous edge points corresponding to the number of its adjacent pixels is large, and the area formed by the continuous edge points corresponding to the number of pixels is a welding area.

[0072] In some embodiments, since the size of the welding area is often smaller than that of the normal area, that is, the number of pixels is smaller, through the second method, the area consisting of continuous edge points corresponding to a smaller number of pixels is more likely to be confirmed as the welding area, and the accuracy of the judgment of the welding area can also be verified by this method.

[0073] The above process has completed the identification of the welding area on the object to be tested. The following needs to further analyze the welding conditions in these welding areas. Generally speaking, the welding quality is judged by the specific information of the welding area, such as whether the height of the weld is flat and consistent, the width of the weld, and the imprint on the surface of the welding area (that is, the part of the welding wire left on the surface of the object during the welding process). These welding characteristics can reflect the quality of welding to a certain extent. When the height of the weld is flat, the width of the weld is similar, and the surface texture of the welding area is good, it means that the welding quality here is high.

[0074] Step 104, determining a welding quality factor according to the height distribution influencing factor of the weld corresponding to the welding area, the weld width performance influencing factor and the texture uniformity.

[0075] The welding quality is judged by comprehensively considering the height distribution of the weld, the width of the weld and the uniformity of the weld texture.

[0076] In some embodiments, step 104 specifically includes:

[0077] Step 1041, determining the height distribution influencing factor of the weld corresponding to the welding area according to the difference between the height of the weld corresponding to the welding area at various locations and the average height.

[0078] The difference between the height of the weld at various locations and the average height can reflect the height distribution of the weld.

[0079] In some embodiments, step 1041 is implemented according to the following formula:

[0080]

[0081] In the formula, Indicates the influence factor of the height distribution of the weld, reflecting the height distribution of the weld. Indicates the side view of the object to be measured. The overall height of the object and weld at that location reflects the total height of the object including the weld at that location. Indicates the side view of the object to be measured. The height of the object at that location reflects the height of the object at that location. Indicates the average height of the weld, reflecting the overall height of the entire weld. represents the position number, and N represents the number of positions. Can reflect the The difference between the weld height at and the average weld height, It can reflect the overall difference between the weld height at various locations and the average weld height, that is, the variance of the weld height, and its value is not zero. is the influence factor of the height distribution of the weld , thus reflecting the height distribution of the weld. The smaller the variance of the weld height, the greater the height distribution influencing factor. The larger it is, the more even the height distribution is, and the more consistent the weld height is, indicating that the weld quality is high in terms of the influence of the weld height distribution.

[0082] like Figure 4 As shown, Indicates the side view of the object to be measured. The overall height of the object and weld, Indicates the side view of the object to be measured. The height of the object itself, Indicates the height of the weld.

[0083] Since welding quality is not only determined by the height of the weld, but also affected by the weld width and texture uniformity of the weld, the present invention further analyzes the weld quality in combination with the weld width and texture uniformity.

[0084] Step 1042, determining a performance influencing factor of a weld bead width of a weld corresponding to the welding area according to the variance of the width of each weld bead corresponding to the welding area and the difference in angles between adjacent edge pixels of the welding area.

[0085] The weld bead width of the weld corresponding to the welding area reflects the uniformity of the weld bead width. The more uniform the weld bead width is, the better the welding quality is.

[0086] In some embodiments, step 1042 is implemented according to the following formula:

[0087]

[0088] In the formula, Indicates the influencing factor of the weld bead width performance of the weld, reflecting the uniformity of the weld bead width performance. The top view area of ​​the weld The difference in angles between adjacent edge pixels reflects the vertical alignment of adjacent pixels. The value of is not zero. The smaller it is, the more aligned the adjacent pixels are in the vertical direction. It represents the variance of the width of the weld bead corresponding to the horizontal direction where all edge pixels are located in the weld top-view area, and its value is not zero, reflecting the change of the width of the weld bead corresponding to the horizontal direction where all edge pixels are located in the weld top-view area. The smaller the value, the smaller the change in weld width. represents the function to find the maximum value, Represents the number of adjacent edge pixels. The influencing factor of the weld width is , thus reflecting the uniformity of the weld width. The smaller, The smaller, The smaller the weld width, the greater the influencing factor The larger it is, the more uniform the width is and the more consistent the weld width is, indicating that the weld quality is high in terms of the weld bead width performance.

[0089] In some embodiments, the difference in angles between adjacent edge pixels of the welding area is determined according to the following formula:

[0090]

[0091] In the formula, The top view area of ​​the weld The difference in angles between adjacent edge pixels reflects the vertical alignment of adjacent pixels. The zth pixel in the weld top view area The angle between the line connecting the edge pixel point and the z-th edge pixel point and the horizontal right direction, represents the angle between the line connecting the zth edge pixel point and the z+1th edge pixel point in the weld top-view area and the horizontal right direction, and z represents the number of edge pixel points in the weld top-view area. The angle between the line connecting the first edge pixel point and the zth edge pixel point and the horizontal right direction, and the angle between the line connecting the zth edge pixel point and the z+1th edge pixel point and the horizontal right direction are the first angles of the weld top view area. The angle between adjacent edge pixels.

[0092] Step 1043, determining the weld shape influencing factor based on the weld width performance influencing factor and the height distribution influencing factor.

[0093] The shape influencing factor for judging the weld shape is obtained by combining the influencing factor of weld width and the influencing factor of height distribution.

[0094] In some embodiments, step 1043 is implemented according to the following formula:

[0095]

[0096] In the formula, represents the influence factor of weld shape, Indicates the influence factor of the height distribution of the weld, Indicates the influencing factor of the weld bead width. The bigger, The larger the The larger it is, the more evenly distributed the height of the weld is and the weld width is even.

[0097] The welding quality is affected by the texture uniformity. The present invention further analyzes the weld quality in combination with the texture uniformity.

[0098] Step 1044, determining the welding quality factor according to the texture uniformity and shape influence factor of the weld corresponding to the welding area.

[0099] The more uniform the texture of the weld is, the better the welding quality is. Introducing the texture uniformity of the weld into the welding quality judgment can make the judgment result more accurate.

[0100] In some embodiments, step 1044 is implemented according to the following formula:

[0101]

[0102] In the formula, Indicates the welding quality factor, reflecting the welding quality. represents the influence factor of weld shape, Indicates the first The pixel value of a pixel, Represents the average value of all pixel values ​​on the feature map corresponding to the weld top view area, Indicates the number of pixels on the feature map corresponding to the weld top view area. Indicates the number of pixels on the feature map corresponding to the weld top view area, represents the normalization function, Not zero. The larger the value is, the smaller the difference in pixel values ​​of all pixels on the feature map corresponding to the weld top view area is. The larger the The larger the normalized The larger the welding quality factor The larger it is, the higher the quality of the weld.

[0103] In some embodiments, a feature map corresponding to the weld top-view area is obtained by a local binary method, and the texture uniformity of the weld corresponding to the welding area is determined based on the pixel value of each pixel point in the feature map.

[0104] Step 105 , in response to the welding quality factor being greater than the quality threshold, the weld corresponding to the welding area passes the quality inspection.

[0105] The welding quality factor is greater than the quality threshold, indicating that the welding quality meets the quality inspection requirements and the weld corresponding to the welding area passes the quality inspection.

[0106] In summary, the present application provides a method for visually inspecting welding quality applied to a welding machine. The method comprehensively determines the welding quality based on the height distribution of the weld, the width of the weld bead and the texture uniformity of the weld, thereby improving the accuracy of quality inspection.

[0107] In the second aspect, the present application provides a welding quality visual inspection system for a welding machine, see Figure 5 The system comprises:

[0108] The acquisition module 501 is used to acquire the welding image.

[0109] The segmentation module 502 is used to perform semantic segmentation processing on the welding image to obtain an object area image and a background area image, wherein the object area image includes the object to be measured, and the background area image includes background information of the photographed object.

[0110] The area detection module 503 is used to perform edge detection on the object area image and determine the welding area according to the change of the gray value of the area formed by continuous edge points, wherein the object area image includes a top view and a side view of the object to be measured, and the welding area includes a top view area and a side view area of ​​the weld on the object to be measured, the top view area of ​​the weld contains the weld bead width information and texture information of the weld, and the side view area of ​​the weld contains the height information of the weld.

[0111] The determination module 504 is used to determine the welding quality factor according to the height distribution influencing factor of the weld corresponding to the welding area, the weld width performance influencing factor and the texture uniformity.

[0112] The quality detection module 505 is used to respond to the weld corresponding to the welding area passing the quality detection in response to the welding quality factor being greater than the quality threshold.

[0113] In some embodiments, the determination module 504 includes:

[0114] A first determining unit is used to determine a height distribution influencing factor of the weld corresponding to the welding area according to the difference between the height of the weld corresponding to the welding area at various locations and the average height;

[0115] A second determination unit is used to determine a performance influencing factor of a weld width of a weld corresponding to the welding area according to a variance of the width of each weld corresponding to the welding area and a difference in angles between adjacent edge pixels of the welding area;

[0116] A third determination unit is used to determine the shape influence factor of the weld based on the weld width performance influence factor and the height distribution influence factor;

[0117] The fourth determining unit is used to determine the welding quality factor according to the texture uniformity and shape influence factor of the weld corresponding to the welding area.

[0118] In some embodiments, the height distribution influencing factor of the weld corresponding to the welding area is determined according to the following formula:

[0119]

[0120] In the formula, Indicates the influence factor of the height distribution of the weld, Indicates the side view of the object to be measured. The overall height of the object and weld, Indicates the side view of the object to be measured. The height of the object itself, represents the average height of the weld, represents the position number, and N represents the number of positions.

[0121] In some embodiments, the weld width of the weld corresponding to the welding area is determined to be influenced by the following factors:

[0122]

[0123] In the formula, Indicates the influencing factor of the weld bead width. The top view area of ​​the weld The difference in angles between adjacent edge pixels, It represents the variance of the width of the weld bead corresponding to all edge pixels in the weld top-view area in the horizontal direction. represents the function to find the maximum value, Represents the logarithm of adjacent edge pixels; It is the maximum value of the difference in angles between adjacent edge pixels in the weld top-view area.

[0124] In some embodiments, the difference in angles between adjacent edge pixels of the welding area is determined according to the following formula:

[0125]

[0126] In the formula, The top view area of ​​the weld The difference in angles between adjacent edge pixels, The zth pixel in the weld top view area The angle between the line connecting the edge pixel point and the z-th edge pixel point and the horizontal right direction, represents the angle between the line connecting the zth edge pixel point and the z+1th edge pixel point in the weld top-view area and the horizontal right direction, and z represents the number of edge pixel points in the weld top-view area. The angle between the line connecting the first edge pixel point and the zth edge pixel point and the horizontal right direction, and the angle between the line connecting the zth edge pixel point and the z+1th edge pixel point and the horizontal right direction are the first angles of the weld top view area. The angle between adjacent edge pixels.

[0127] In some embodiments, the shape influence factor of the weld is determined according to the following formula:

[0128]

[0129] In the formula, represents the influence factor of weld shape, Indicates the influence factor of the height distribution of the weld, Indicates the influencing factor of the weld bead width.

[0130] In some embodiments, the welding quality factor is determined according to the following formula:

[0131]

[0132] In the formula, represents the welding quality factor, represents the influence factor of weld shape, Indicates the first The pixel value of a pixel, Represents the average value of all pixel values ​​on the feature map corresponding to the weld top view area, Indicates the number of pixels on the feature map corresponding to the weld top view area. Indicates the number of pixels on the feature map corresponding to the weld top view area, Represents the normalization function.

[0133] In some embodiments, a feature map corresponding to the weld top-view area is obtained by a local binary method, and the texture uniformity of the weld corresponding to the welding area is determined based on the pixel value of each pixel point in the feature map.

[0134] In some embodiments, the region detection module 503 includes:

[0135] An edge detection unit is used to perform edge detection on the object area image to obtain an edge detection result;

[0136] A fifth determining unit, used to determine the grayscale value of the area formed by the continuous edge points in the edge detection result, and obtain a grayscale value sequence of the area formed by the continuous edge points;

[0137] The sixth determination unit is used to determine that, in response to the difference between a certain grayscale value in the grayscale value sequence and its adjacent grayscale values ​​being greater than a first grayscale difference threshold, an area formed by continuous edge points corresponding to the grayscale value is a welding area, and the number of grayscale values ​​whose differences with its adjacent grayscale values ​​are greater than the first grayscale difference threshold is the number of welding areas, or, sorting the number of pixels in the area formed by continuous edge points in order from small to large to obtain a pixel number sorting series of the area formed by continuous edge points, and in response to the difference in grayscale values ​​of the area formed by continuous edge points corresponding to the number of pixels in the pixel number sorting series and the number of adjacent pixels being greater than a second grayscale difference threshold, the area formed by continuous edge points corresponding to the grayscale value is determined to be a welding area.

[0138] In summary, the present application provides a welding quality visual inspection system for a welding machine, which judges the welding quality based on the height distribution of the weld, the width of the weld and the texture uniformity of the weld, thereby improving the accuracy of quality inspection.

[0139] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0140] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for visually inspecting welding quality applied to a welding machine, characterized in that: The method comprises: Acquire welding images; Performing semantic segmentation processing on the welding image to obtain an object area image and a background area image, wherein the object area image includes the object to be measured, and the background area image includes background information of the photographed object; Performing edge detection on the object area image, and determining the welding area according to the gray value change of the area formed by continuous edge points, wherein the object area image includes a top view and a side view of the object to be measured, and the welding area includes a top view area and a side view area of ​​a weld on the object to be measured, the top view area of ​​the weld contains weld bead width information and texture information of the weld, and the side view area of ​​the weld contains height information of the weld; Determine the welding quality factor according to the height distribution influencing factor of the weld corresponding to the welding area, the weld width performance influencing factor and the texture uniformity; In response to the welding quality factor being greater than a quality threshold, the weld corresponding to the welding area passes the quality inspection; The method for obtaining the welding quality factor is as follows: according to the difference between the height of the weld corresponding to the welding area at various locations and the average height, the height distribution influence factor of the weld corresponding to the welding area is determined; according to the variance of the width of the weld corresponding to the welding area at various locations and the difference in angles between adjacent edge pixels of the welding area, the weld width performance influence factor of the weld corresponding to the welding area is determined; based on the weld width performance influence factor and the height distribution influence factor, the shape influence factor of the weld is determined; according to the texture uniformity and shape influence factor of the weld corresponding to the welding area, the welding quality factor is determined; Among them, the influencing factor of the weld width of the weld corresponding to the welding area is determined: ， In the formula, Indicates the influencing factor of the weld bead width. The top view area of ​​the weld The difference in angles between adjacent edge pixels, Indicates the difference in angle between the first pair of adjacent edge pixels in the weld top view area. Indicates the difference in angle between the second pair of adjacent edge pixels in the weld top view area. Indicates the difference in angle between the third pair of adjacent edge pixels in the weld top view area. It represents the variance of the width of the weld bead corresponding to all edge pixels in the weld top-view area in the horizontal direction. represents the function to find the maximum value, Represents the logarithm of adjacent edge pixels; is the maximum value of the difference in angles between adjacent edge pixels in the weld top-view area; Among them, a feature map corresponding to the top-view area of ​​the weld is obtained by a local binary method, and the texture uniformity of the weld corresponding to the welding area is determined based on the pixel value of each pixel point in the feature map.

2. The welding quality visual inspection method applied to a welding machine according to claim 1 is characterized in that: According to the following formula, the height distribution influencing factor of the weld corresponding to the welding area is determined: ， In the formula, Indicates the influence factor of the height distribution of the weld, Indicates the side view of the object to be measured. The overall height of the object and weld at Indicates the side view of the object to be measured. The height of the object itself, Indicates the average height of the weld, represents the position number, and N represents the number of positions.

3. The welding quality visual inspection method applied to a welding machine according to claim 1 is characterized in that: The difference in angles between adjacent edge pixels of the welding area is determined according to the following formula: ， In the formula, The top view area of ​​the weld The difference in angles between adjacent edge pixels, The zth pixel in the weld top view area The angle between the line connecting the edge pixel point and the z-th edge pixel point and the horizontal right direction, It represents the angle between the line between the zth edge pixel point and the z+1th edge pixel point in the weld top-view area and the horizontal right direction, where z represents the number of edge pixels in the weld top-view area.

4. The welding quality visual inspection method applied to a welding machine according to claim 1 is characterized in that: According to the following formula, the shape influencing factor of the weld is determined: ， In the formula, represents the influence factor of weld shape, Indicates the influence factor of the height distribution of the weld, Indicates the influencing factor of the weld bead width.

5. The welding quality visual inspection method applied to a welding machine according to claim 1 is characterized in that: The welding quality factor is determined according to the following formula: ， In the formula, represents the welding quality factor, represents the influence factor of weld shape, Indicates the first The pixel value of a pixel, Represents the average value of all pixel values ​​on the feature map corresponding to the weld top view area, Indicates the number of pixels on the feature map corresponding to the weld top view area. Indicates the number of pixels on the feature map corresponding to the weld top view area, Represents the normalization function.

6. The welding quality visual inspection method applied to a welding machine according to claim 1 is characterized in that: Performing edge detection on the object area image and determining the welding area according to the gray value change of the area formed by continuous edge points includes: Performing edge detection on the object area image to obtain an edge detection result; In the edge detection result, the grayscale value of the area formed by the continuous edge points is determined to obtain a grayscale value sequence of the area formed by the continuous edge points; In response to the fact that the difference between a certain grayscale value in the grayscale value sequence and its adjacent grayscale values ​​is greater than a first grayscale difference threshold, the area formed by the continuous edge points corresponding to the grayscale value is determined to be the welding area, and the number of grayscale values ​​whose differences with its adjacent grayscale values ​​are greater than the first grayscale difference threshold is the number of welding areas, or, the number of pixels in the area formed by the continuous edge points is sorted in order from small to large to obtain a pixel number sorting series of the area formed by the continuous edge points, and in response to the fact that the difference in grayscale values ​​of the area formed by the continuous edge points corresponding to the number of pixels in the pixel number sorting series and the number of adjacent pixels is greater than a second grayscale difference threshold, the area formed by the continuous edge points corresponding to the grayscale value is determined to be the welding area.

7. A welding quality visual inspection system for welding machines, characterized in that: The system comprises: An acquisition module, used for acquiring welding images; A segmentation module, used for performing semantic segmentation processing on the welding image to obtain an object area image and a background area image, wherein the object area image includes the object to be measured, and the background area image includes background information of the photographed object; A region detection module is used to perform edge detection on the object region image, and determine the welding region according to the gray value change of the region formed by continuous edge points, wherein the object region image includes a top view and a side view of the object to be measured, and the welding region includes a top view region and a side view region of a weld on the object to be measured, the top view region of the weld contains weld bead width information and texture information of the weld, and the side view region of the weld contains height information of the weld; A determination module, used to determine a welding quality factor according to a height distribution influencing factor of a weld corresponding to the welding area, a weld width performance influencing factor and a texture uniformity; The method for obtaining the welding quality factor is as follows: according to the difference between the height of the weld corresponding to the welding area at various locations and the average height, the height distribution influence factor of the weld corresponding to the welding area is determined; according to the variance of the width of the weld corresponding to the welding area at various locations and the difference in angles between adjacent edge pixels of the welding area, the weld width performance influence factor of the weld corresponding to the welding area is determined; based on the weld width performance influence factor and the height distribution influence factor, the shape influence factor of the weld is determined; according to the texture uniformity and shape influence factor of the weld corresponding to the welding area, the welding quality factor is determined; Among them, the influencing factor of the weld width of the weld corresponding to the welding area is determined: ， In the formula, Indicates the influencing factor of the weld bead width. The top view area of ​​the weld The difference in angles between adjacent edge pixels, Indicates the difference in angle between the first pair of adjacent edge pixels in the weld top view area. Indicates the difference in angle between the second pair of adjacent edge pixels in the weld top view area. Indicates the difference in angle between the third pair of adjacent edge pixels in the weld top view area. It represents the variance of the width of the weld bead corresponding to all edge pixels in the weld top-view area in the horizontal direction. represents the function to find the maximum value, Represents the logarithm of adjacent edge pixels; is the maximum value of the difference in angles between adjacent edge pixels in the weld top-view area; The feature map corresponding to the weld top-view area is obtained by a local binary method, and the texture uniformity of the weld corresponding to the weld area is determined based on the pixel value of each pixel point in the feature map; The quality detection module is used for responding that the welding quality factor is greater than the quality threshold and the weld corresponding to the welding area passes the quality detection.

Images