A method for detecting the quality of a welded steel structure of a construction engineering

Abstract

The application relates to the technical field of welded steel structure quality detection, in particular to a welded steel structure quality detection method for building engineering, which comprises the following steps: acquiring a gray-scale image of a welded steel structure area; presetting each straight line to divide the welded steel structure area, and screening symmetrical lines from all the straight lines; acquiring growth stop pixel points of each pixel point on the symmetrical lines on both sides of the symmetrical lines; acquiring a welding seam area and a characteristic length thereof through the connecting lines of the growth stop pixel points of the pixel points; dividing the welding seam area into each local welding seam area, acquiring a gray-scale difference degree of each pixel point in each local welding seam area; acquiring a texture characteristic value of each local welding seam area; acquiring a texture approximation degree between any two local welding seam areas; acquiring a welding quality evaluation value of the welded steel structure, and detecting the quality of the welded steel structure. The application aims to improve the precision of the quality detection of the welded steel structure.

Description

Technical Field

[0001] This application relates to the field of welded steel structure quality inspection technology, specifically to a method for inspecting the quality of welded steel structures in building engineering. Background Technology

[0002] Welded steel structures in building construction are engineering forms that connect steel materials (such as structural steel and steel plates) into load-bearing or supporting structures through welding processes. They are characterized by high strength, high rigidity, and fast construction speed, and their quality directly affects the safety and durability of buildings. Traditional welding quality inspection methods mainly rely on manual visual inspection or non-destructive testing techniques such as ultrasound and X-rays, which suffer from low efficiency, high cost, and susceptibility to subjective factors. In recent years, image processing-based inspection methods have gradually emerged, but existing technologies usually rely on fixed threshold segmentation or simple texture analysis, which is difficult to adapt to the complexity and diversity of welded steel structures, resulting in low quality inspection accuracy. Summary of the Invention

[0003] In view of the above, it is necessary to provide a machine vision-based method for detecting color difference in the coating of large-diameter pressure steel pipes, which improves the accuracy of quality inspection of welded steel structures compared with traditional methods.

[0004] The present application discloses a method for quality inspection of welded steel structures in building engineering, which adopts the following technical solution:

[0005] One embodiment of this application provides a method for quality inspection of welded steel structures in building engineering, the method comprising the following steps:

[0006] Obtain the grayscale image of the welded steel structure area;

[0007] The welded steel structure area is divided by pre-defined straight lines. Symmetrical lines are selected from all straight lines based on the similarity of pixel distribution in the two segmented areas of each straight line.

[0008] By identifying the grayscale difference between each pixel on the symmetry line and the pixel in the direction perpendicular to the symmetry line, the growth-stopping pixels on both sides of the symmetry line are obtained; by connecting the growth-stopping pixels, the weld area and its feature length are obtained.

[0009] The weld area is divided into local weld areas by the pixels on the symmetry line. The grayscale difference of each pixel is obtained by the grayscale difference between each pixel in each local weld area and its preset nearest neighbor pixels.

[0010] By analyzing the distribution of grayscale differences among pixels within each local weld area, and comparing the length of the line connecting each pixel on the symmetry line within each local weld area to its growth-stopping pixel with the feature length, the texture feature value of each local weld area is obtained.

[0011] For any two local weld seam regions, the texture approximation between the two local weld seam regions is obtained by comparing the distances between pixels at the same position on the symmetry line and the growth-stopping pixels on the same side of the symmetry line, and by comparing the texture feature values.

[0012] By using the texture feature values, as well as the average level and dispersion of the texture similarity between each local weld area and all other local weld areas, a welding quality assessment value for the welded steel structure is obtained, which is used to detect the quality of the welded steel structure.

[0013] In one embodiment, the method for selecting symmetrical lines from all straight lines is as follows:

[0014] Obtain the minimum bounding rectangle of the welded steel structure area, and draw a preset number of straight lines through the center point of the minimum bounding rectangle to obtain the edge pixels within the welded steel structure area;

[0015] The two regions obtained by dividing the area by any straight line are denoted as the first region and the second region, respectively. The line connecting any edge pixel in the first region and the center point is obtained. The edge pixel in the second region that passes through the extension of the line is taken as the symmetrical point of the edge pixel.

[0016] Calculate the difference in the number of edge pixels between the first region and the second region. Record the distances between each edge pixel in the first region, the symmetrical point of each edge pixel, and the center point as the first distance and the second distance, respectively. Calculate the sum of the differences between the first distance and the second distance for all edge pixels in the first region.

[0017] Calculate the inverse proportional mapping result of the product of the difference and the sum; take the line with the largest inverse proportional mapping result as the line of symmetry.

[0018] In one embodiment, the process of obtaining the growth-stopping pixel is as follows:

[0019] For each pixel, growth is performed in both directions perpendicular to the line of symmetry. Each pixel is taken as an initial growth pixel. The growth condition is: calculate the grayscale difference between the growth pixel and its neighboring pixels. The ratio of the difference value to the grayscale value of the growth pixel is less than or equal to a preset threshold. The last pixel obtained by growth in each direction is taken as the growth stop pixel.

[0020] In one embodiment, the process of obtaining the weld region and its characteristic length is as follows:

[0021] Obtain the connecting line between two growth-stopping pixels of each pixel, and form a weld area by connecting the connecting lines of all pixels on the symmetry line; count the length of the connecting line of all pixels on the symmetry line that appears most frequently, and use it as the feature length.

[0022] In one embodiment, the method for dividing the local weld area is as follows:

[0023] Each segmentation pixel is evenly preset on the line of symmetry, and the weld area is divided into local weld areas by using a straight line passing through the segmentation pixel and perpendicular to the line of symmetry.

[0024] In one embodiment, the process of obtaining the grayscale difference is as follows:

[0025] The maximum grayscale difference between each pixel and its preset nearest neighbor pixels along the symmetry line direction is calculated; the maximum grayscale difference between each pixel and its preset nearest neighbor pixels along the direction perpendicular to the symmetry line direction is also calculated.

[0026] The grayscale difference is directly proportional to the maximum value and inversely proportional to the maximum value.

[0027] In one embodiment, the expression for the texture feature value is:

[0028] In the formula, Z m N represents the texture feature value of the m-th local weld region; m T represents the number of pixels within the m-th local weld region; m,n Let d represent the grayscale difference of the nth pixel within the m-th local weld region; calculate the minimum deviation of the grayscale difference between the nth pixel and all other pixels within the m-th local weld region. m,n σ represents the distance between the nth pixel in the mth local weld area and the pixel corresponding to the minimum value; σ represents a preset positive number; F represents the number of pixels on the symmetry line in a single local weld area; The length of the line connecting the f-th pixel on the symmetry line in the m-th local weld region and the growth-stopping pixel on the i-th side of the symmetry line is represented by D; the feature length is represented by |*|; and the absolute value operation is represented by |*|.

[0029] In one embodiment, the expression for the texture approximation is:

[0030] In the formula, U A,BThis indicates the textural similarity between local weld region A and local weld region B; exp() represents an exponential function with the natural constant as its base; Z A Z B These represent the texture feature values ​​of local weld area A and local weld area B, respectively; F represents the number of pixels on the symmetry line within a single local weld area. These represent the lengths of the lines connecting the f-th pixel on the line of symmetry in local weld regions A and B to the pixel on the i-th side of the line of symmetry where growth has stopped; |*| represents the absolute value operation.

[0031] In one embodiment, the process of obtaining the welding quality assessment value is as follows:

[0032] Calculate the mean of the texture similarity between each local weld area and all other local weld areas;

[0033] Calculate the product of the texture feature value and the mean;

[0034] Map the dispersion to a positive value, and calculate the ratio of the product to the positive value;

[0035] The welding quality assessment value is the normalized value of the sum of the ratios of all local weld areas.

[0036] In one embodiment, the method for detecting the quality of the welded steel structure is as follows:

[0037] When the welding quality assessment value is greater than or equal to the preset welding quality assessment value threshold, the welded steel structure is deemed to be of qualified quality; otherwise, it is deemed to be unqualified.

[0038] This application has at least the following beneficial effects:

[0039] This application acquires distortion-free, low-reflectivity images of the welding area using a high-resolution industrial camera combined with a polarizing filter; by utilizing the symmetry characteristics of the welded steel structure, symmetry lines are selected through segmentation and similarity analysis, which can accurately locate the weld area and provide a key reference for subsequent weld inspection; by obtaining growth-stopping pixels, the boundary of the weld area can be precisely defined; by statistically analyzing the frequency of occurrence of the line lengths connecting the growth-stopping pixels, the characteristic length of the weld area is obtained, providing an important quantitative indicator for subsequent welding quality assessment.

[0040] Furthermore, dividing the weld area into multiple local weld areas allows for detailed analysis of the local features of the weld, which helps in the detection of local welding defects. By obtaining grayscale differences through the local texture characteristics of the weld area, changes in welding quality can be more sensitively reflected, improving the sensitivity of quality inspection. By calculating texture feature values, the texture characteristics of the weld area can be effectively reflected, and the presence of welding defects can be effectively detected, which helps improve the accuracy of subsequent quality inspection.

[0041] Furthermore, the calculation of texture approximation considers the degree of similarity in shape and texture features between local welded areas, which can effectively detect differences in local weld quality, provide a basis for assessing the uniformity of weld quality, and enable more accurate assessment of weld quality, thereby improving detection precision. By comprehensively considering texture characteristic values, the average level and dispersion of texture approximation, the quality of welded steel structures can be comprehensively assessed, providing a comprehensive quality index. This quantifies the weld quality assessment index into a specific numerical value, facilitating comparison with preset thresholds, achieving automated quality inspection, and improving the accuracy of quality inspection of welded steel structures. Attached Figure Description

[0042] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 A flowchart illustrating the steps of a method for quality inspection of welded steel structures in building engineering, provided in this application;

[0044] Figure 2 This is a schematic diagram illustrating the process of obtaining the line of symmetry.

[0045] Figure 3 This is a schematic diagram illustrating the process of obtaining welding quality assessment values. Detailed Implementation

[0046] In the description of the embodiments in this application, the words "exemplary," "or," and "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplary," "or," and "for example" is intended to present the relevant concepts in a specific manner.

[0047] Unless otherwise defined, 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 application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. It should be understood that, unless otherwise stated, " / " in this application means "or".

[0048] It should also be noted that the terms "first" and "second" in this application are used to distinguish similar objects, rather than to describe a specific order or sequence.

[0049] The following description, in conjunction with the accompanying drawings, details a specific scheme for a method of inspecting the quality of welded steel structures in building engineering provided in this application.

[0050] This application provides an embodiment of a method for quality inspection of welded steel structures in building engineering. Specifically, it provides the following method for quality inspection of welded steel structures in building engineering. Please refer to [link / reference]. Figure 1 The method includes the following steps:

[0051] Step 1: Obtain the grayscale image of the welded steel structure area.

[0052] Surface images of the welded steel structure are acquired using a high-resolution industrial camera. An LED ring light is employed to provide uniform illumination, and a polarizing filter is used to reduce glare interference from the metal surface. The camera is positioned perpendicular to the welded steel structure, i.e., shooting from above, to ensure distortion-free images. The acquired surface images of the welded steel structure are RGB images, which are then converted to grayscale to obtain grayscale images. Grayscale conversion is a well-known technique and will not be elaborated upon in this application.

[0053] Furthermore, the grayscale image of the welded steel structure surface is processed by a threshold segmentation algorithm to divide the grayscale image of the welded steel structure surface into a foreground and a background, wherein the foreground is the grayscale image of the welded steel structure area.

[0054] In this embodiment, the threshold segmentation algorithm is the Otsu threshold segmentation algorithm. The Otsu threshold segmentation algorithm is a well-known technology and will not be described in detail in this application. As other implementation methods, based on the ability to obtain the grayscale image of the welded steel structure area, the implementer may adopt other existing technologies, such as iterative threshold segmentation, etc. This application does not impose any special restrictions.

[0055] Step 2: Preset each straight line to divide the welded steel structure area. Based on the similarity of the distribution of pixels in the two segmented areas in the segmentation results of each straight line, select symmetrical lines from all the straight lines.

[0056] Since welded steel structures in building construction are formed by joining and welding two steel pieces together, the overall structure of the steel structure is symmetrical in terms of the welded steel surface image, and the weld area is located on the line of symmetry of the steel structure. Therefore, based on the above characteristics, the weld area is extracted.

[0057] Obtain the minimum bounding rectangle of the welded steel structure region. Draw a predetermined number of straight lines through the center point of the minimum bounding rectangle to divide the welded steel structure region. The included angle between any two adjacent straight lines is equal. Obtain each edge pixel point within the welded steel structure region. The method for obtaining the minimum bounding rectangle is a known technique and will not be elaborated upon in this application.

[0058] In this embodiment, the preset quantity is 20, and the included angle between any two adjacent straight lines is 18°. The preset quantity is preset by the user and can be set by the implementer. This application does not impose any special restrictions.

[0059] In this embodiment, the Roberts operator is used to obtain edge pixels. The Roberts operator is a well-known technology and will not be described in detail here. As other implementation methods, based on the ability to obtain edge pixels, implementers may use other existing technologies, such as the Prewitt operator, the Sobel operator, etc. This application does not impose any special restrictions.

[0060] Furthermore, based on the similarity of pixel distribution within two segmented regions in the segmentation results of each straight line, symmetrical lines are selected from all straight lines, specifically as follows:

[0061] Divide the welded steel structure area into two regions by any straight line, and denote them as the first region and the second region, respectively. Obtain the line connecting any edge pixel in the first region to the center point. Take the edge pixel in the second region that passes through the extension of the line as the symmetrical point of the edge pixel. When there is more than one edge pixel in the second region that passes through the extension of the line, denote the length of the line connecting any edge pixel to the center line as the reference length. Denote the lengths of the lines connecting each edge pixel in the second region that passes through the extension of the line to the center point as the comparison lengths. Take the edge pixel corresponding to the comparison length with the smallest difference from the reference length as the symmetrical point of the pixel.

[0062] The segmentation excellence of any straight line is obtained by considering the difference in the number of edge pixels between the first region and the second region, as well as the difference in distance from any edge pixel and its symmetrical point to the center point. The expression is as follows:

[0063] In the formula, R jThe segmentation quality of the j-th line is represented by exp(); exp() represents an exponential function with the natural constant as the base; P j,1 P j,2 d represents the number of edge pixels in the first and second regions obtained by dividing the welded steel structure region with the j-th straight line, respectively; p ,d' p and represent the distance between the p-th edge pixel and the center point within the first region segmented by the j-th straight line, and the distance between the symmetrical point of the p-th edge pixel and the center point, respectively; |*| represents the absolute value operation. The exponential function with the natural constant as the base is merely one embodiment of this application. In another embodiment, the natural constant can be replaced with 3. Provided it is greater than 1, the implementer can replace the natural constant with other values; this application does not impose any special restrictions. Let d p ,d' p These are denoted as the first distance and the second distance, respectively.

[0064] In this embodiment, the difference between the reference length and the comparison length is the absolute value of the difference.

[0065] In this embodiment, the distance between the edge pixel and the center point is the Euclidean distance, and the distance between the symmetrical point of the edge pixel and the center point is the Euclidean distance.

[0066] It should be noted that: the greater the distance difference, the worse the symmetry between the edge pixel and its symmetrical point, and the lower the segmentation quality; the greater the segmentation quality, the more consistent the two regions obtained by the i-th straight line segmenting the welded steel structure area, that is, the more symmetrical the two regions, and the more likely the i-th straight line is to be a symmetrical line.

[0067] Furthermore, the line with the highest segmentation quality is taken as the line of symmetry. A schematic diagram of the process for obtaining the line of symmetry is shown below. Figure 2 As shown.

[0068] Step 3: Obtain the growth-stopping pixels on both sides of the symmetry line by the grayscale difference between each pixel on the symmetry line and the pixels in the direction perpendicular to the symmetry line; obtain the weld area and its feature length by connecting the growth-stopping pixels.

[0069] Since welded steel structures are made by joining two identical steel pieces, the welding process is typically linear, resulting in a striped, regular texture in the weld area. This texture is usually perpendicular to the line of symmetry. The grayscale values ​​of pixels within the weld area often differ from those on the steel surface, while the texture within the weld area exhibits a striped distribution, with the grayscale values ​​of pixels on these stripes being relatively similar.

[0070] Based on the above analysis, for each pixel on the symmetry line, the pixel is grown in both directions perpendicular to the symmetry line. The pixel is taken as the initial growth pixel. The growth condition is: calculate the gray value difference between the growth pixel and its neighboring pixels. The ratio of the difference value to the gray value of the growth pixel is less than or equal to a preset threshold. The last pixel obtained by the growth on both sides of the symmetry line is taken as the growth stop pixel on both sides of the symmetry line.

[0071] The preset threshold value is 0.3. The preset threshold value is preset by the user and can be set by the implementer. This application does not impose any special restrictions.

[0072] Furthermore, the connecting lines between two growth-stopping pixels of each pixel on the symmetry line are obtained, and the connecting lines of all pixels on the symmetry line are used to form a weld area; the length with the highest frequency among the lengths of the connecting lines of all pixels on the symmetry line is counted and used as the characteristic length of the weld area, that is, the gap length between the two steel parts.

[0073] Step 4: Divide the weld area into local weld areas by using the pixels on the symmetry line. Obtain the grayscale difference of each pixel in each local weld area and its preset nearest neighbor pixels by using the grayscale difference between each pixel in each local weld area and its preset nearest neighbor pixels.

[0074] In actual welding, excessive welding material or spatter may cause the actual weld area to be larger than the preset weld area, resulting in poor welding quality. Therefore, the welding quality can be reflected by the difference between the length of the line connecting each pixel and its growing pixels and the feature length.

[0075] Since regular texture information usually exists within the weld seam area, this embodiment divides the weld seam area into local weld seam areas using pixels along the symmetry line. Specifically, segmentation pixels are evenly preset along the symmetry line, and the weld seam area is divided into local weld seam areas using straight lines passing through the segmentation pixels and perpendicular to the symmetry line. This facilitates subsequent comparison of the features of different local areas within the weld seam area.

[0076] In this embodiment, the number of local weld seam regions obtained by segmentation is 20. The number of local weld seam regions is preset by human intervention and can be set by the implementer. This application does not impose any special restrictions.

[0077] Based on the texture characteristics of the weld seam area, it can be seen that the local grayscale distribution of the pixels should satisfy the texture feature that the grayscale value difference is large in the direction of the symmetry line and small in the direction perpendicular to the symmetry line.

[0078] Based on the above analysis, the grayscale difference of each pixel in each local weld area is obtained by the grayscale difference between each pixel in each local weld area and its preset nearest neighbor pixels, and the expression is:

[0079] In the formula, T m,n C represents the grayscale difference of the nth pixel within the mth local weld area; m,n1 C represents the maximum grayscale difference between the nth pixel within the m-th weld region and its preset nearest neighbor pixels along the symmetry line direction; m,n2 ε represents the maximum grayscale difference between the nth pixel in the mth weld area and its preset nearest neighbor pixels in the direction perpendicular to the line of symmetry; ε represents a preset positive number to avoid a denominator of 0. The value of ε is preset by the user and can be set by the implementer. In this implementation, the value of ε is 0.01.

[0080] In this embodiment, the nearest neighbor pixels of the nth pixel in the direction of the symmetry line are: the two adjacent pixels of the nth pixel in the direction of the symmetry line; the nearest neighbor pixels of the nth pixel in the direction perpendicular to the symmetry line are: the two adjacent pixels of the nth pixel in the direction perpendicular to the symmetry line.

[0081] In this embodiment, the grayscale differences involved in the calculation of grayscale differences are all absolute values ​​of the differences between grayscale values.

[0082] It should be noted that: the greater the grayscale difference, the more the local grayscale distribution of the pixels in the m-th local weld area conforms to the characteristics of linear welding, and the lower the probability of defects appearing in the weld area; the smaller the grayscale difference, the less the local grayscale distribution of the pixels in the m-th local weld area conforms to the characteristics of linear welding, and the higher the probability of defects appearing in the weld area.

[0083] Step 5: By analyzing the distribution of grayscale differences among pixels within each local weld area, and comparing the length of the line connecting each pixel on the symmetry line within each local weld area to its growth-stopping pixel with the feature length, the texture feature value of each local weld area is obtained.

[0084] Since the texture of the weld area in linear welding is usually symmetrical, the texture feature value of each local weld area is obtained by analyzing the distribution of grayscale differences among pixels within each local weld area and comparing the length of the line connecting each pixel on the symmetry line within each local weld area to its growth-stopping pixel with the feature length. The expression is as follows:

[0085] In the formula, Z m N represents the texture feature value of the m-th local weld region;m T represents the number of pixels within the m-th local weld region; m,n Let d represent the grayscale difference of the nth pixel within the m-th local weld region; calculate the minimum deviation of the grayscale difference between the nth pixel and all other pixels within the m-th local weld region. m,n σ represents the distance between the nth pixel in the mth local weld area and the pixel corresponding to the minimum value; σ represents a preset positive number to avoid the denominator being 0. The value of σ is preset by the user and can be set by the implementer. In this implementation, the value of σ is 0.01; F represents the number of pixels on the symmetry line in a single local weld area. The length of the line connecting the f-th pixel on the symmetry line in the m-th local weld region and the growth-stopping pixel on the i-th side of the symmetry line is represented by D; the feature length is represented by |*|; and the absolute value operation is represented by |*|.

[0086] It should be noted that: d m,n The larger the value, the more isolated the local welding characteristics of the nth pixel are, the greater the possibility of noise interference, and the smaller the calculated texture feature value. The larger the texture feature value, the more the texture distribution in the mth local weld area conforms to the texture characteristics of welding, the less the texture in the mth local weld area is damaged, and the lower the possibility of defects. The smaller the texture feature value, the less the texture distribution in the mth local weld area conforms to the texture characteristics of welding, the greater the degree of texture damage in the mth local weld area, and the higher the possibility of defects.

[0087] Step 6: For any two local weld seam areas, the texture approximation between the two local weld seam areas is obtained by comparing the distances between pixels at the same position on the symmetry line and the growth-stopping pixels on the same side of the symmetry line, and by comparing the texture feature values.

[0088] Since the texture of each weld area is regularly distributed in linear welding, for any two local weld areas, the texture approximation between them is obtained by comparing the distances between pixels at the same position on the symmetry line and their growth-stopping pixels on the same side of the symmetry line, and by comparing the texture feature values. The expression is as follows:

[0089] In the formula, U A,B This indicates the textural similarity between local weld region A and local weld region B; exp() represents an exponential function with the natural constant as its base; Z A Z BThese represent the texture feature values ​​of local weld area A and local weld area B, respectively; F represents the number of pixels on the symmetry line within a single local weld area. These represent the lengths of the lines connecting the f-th pixel on the line of symmetry in local weld regions A and B to the pixel on the i-th side of the line of symmetry where growth has stopped; |*| represents the absolute value operation.

[0090] It should be noted that: This reflects the similarity in shape between local weld area A and local weld area B.

[0091] Step 7: Obtain the welding quality assessment value of the welded steel structure by using the texture feature values, as well as the average level and dispersion of the texture similarity between each local weld area and all other local weld areas, for the purpose of detecting the quality of the welded steel structure.

[0092] Furthermore, the welding quality assessment value of the welded steel structure is obtained through the texture feature values, and the average level and dispersion of the texture similarity between each local weld area and all other local weld areas. The expression is as follows:

[0093] In the formula, ZL represents the welding quality assessment value of the welded steel structure; norm() represents the normalization function; M represents the number of local weld areas; Z m This represents the texture feature value of the m-th local weld region; SZ represents the mean of the texture similarity between the m-th local weld region and all other local weld regions. m This represents the dispersion of the texture approximation between the m-th local weld region and all other local weld regions.

[0094] In this embodiment, the dispersion is information entropy. The calculation of information entropy is a well-known technique and will not be described in detail here. As other implementation methods, based on the ability to measure the unevenness of the distribution of texture approximation, the implementer may use other existing techniques, such as variance, coefficient of variation, etc., which will not be described in detail here.

[0095] It should be noted that: a higher mean value for texture similarity indicates less quality fluctuation in local welded areas and higher overall weld quality; a higher dispersion indicates greater weld quality fluctuation in the weld area and a greater likelihood of defects within the weld area. A schematic diagram illustrating the process of obtaining weld quality assessment values ​​is shown below. Figure 3 As shown.

[0096] Furthermore, a welding quality assessment threshold is set. When the welding quality assessment value of the weld area is greater than or equal to the welding quality assessment threshold, the welded steel structure is deemed to be of qualified quality and can be shipped from the factory. Otherwise, the welded steel structure is deemed to be of unqualified quality and needs to be re-welded.

[0097] In this embodiment, the welding quality assessment threshold is set to 0.7. The welding quality assessment threshold is preset by the user and can be set by the implementer according to the actual situation. This application does not impose any special restrictions.

[0098] In summary, this application acquires distortion-free, low-reflectivity welding area images using a high-resolution industrial camera combined with a polarizing filter; by utilizing the symmetry features of the welded steel structure and selecting symmetry lines through segmentation and similarity analysis, the weld area can be accurately located, providing a crucial reference for subsequent weld inspection; by acquiring growth-stopping pixels, the boundary of the weld area can be precisely defined; and by statistically analyzing the frequency of occurrence of the connection lengths of growth-stopping pixels, the characteristic length of the weld area is obtained, providing an important quantitative indicator for subsequent welding quality assessment.

[0099] Furthermore, dividing the weld area into multiple local weld areas allows for detailed analysis of the local features of the weld, which helps in the detection of local welding defects. By obtaining grayscale differences through the local texture characteristics of the weld area, changes in welding quality can be more sensitively reflected, improving the sensitivity of quality inspection. By calculating texture feature values, the texture characteristics of the weld area can be effectively reflected, and the presence of welding defects can be effectively detected, which helps improve the accuracy of subsequent quality inspection.

[0100] Furthermore, the calculation of texture approximation considers the degree of similarity in shape and texture features between local welded areas, which can effectively detect differences in local weld quality, provide a basis for assessing the uniformity of weld quality, and enable more accurate assessment of weld quality, thereby improving detection precision. By comprehensively considering texture characteristic values, the average level and dispersion of texture approximation, the quality of welded steel structures can be comprehensively assessed, providing a comprehensive quality index. This quantifies the weld quality assessment index into a specific numerical value, facilitating comparison with preset thresholds, achieving automated quality inspection, and improving the accuracy of quality inspection of welded steel structures.

[0101] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than that shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. In the descriptions corresponding to the flowcharts and block diagrams in the accompanying drawings, the operations or steps corresponding to different blocks may also occur in a different order than disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. Each block in a block diagram and / or flowchart, and combinations of blocks in a block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0102] It will be apparent to those skilled in the art that this application is not limited to the details of the exemplary embodiments described above, and that this application can be implemented in other specific forms without departing from its essential characteristics. Therefore, the embodiments described above should be considered exemplary and non-limiting in all respects.

Claims

1. A method for quality control of welded steel structures of construction engineering, characterized in that, The method comprises the following steps: Obtain a gray image of a welded steel structure region; Pre-set each straight line to divide the welded steel structure region, and select a symmetry line from all straight lines by the similarity of the distribution of the pixel points in the two divided regions of each straight line; Obtain the growth stop pixel points of each pixel point on the symmetry line by the gray difference between each pixel point and the pixel point perpendicular to the symmetry line, and obtain the weld region and its characteristic length by the connecting line of the growth stop pixel points of each pixel point; Divide the weld region into each local weld region by the pixel points on the symmetry line, and obtain the gray difference degree of each pixel point by the gray difference between each pixel point and its preset neighbor pixel point in each local weld region; Obtain the texture feature value of each local weld region by the distribution of the gray difference degree of the pixel points in each local weld region, and compare the length of the connecting line between each pixel point on the symmetry line and its growth stop pixel point with the characteristic length; Obtain the texture approximation degree between any two local weld regions by comparing the distance between the pixel points at the same position on the symmetry line and the growth stop pixel points on the same side of the symmetry line, and comparing the texture feature values; Obtain the welding quality evaluation value of the welded steel structure by the texture feature value, the average level and the dispersion of the texture approximation degree between each local weld region and all other local weld regions, which is used to detect the quality of the welded steel structure; The method for selecting a symmetry line from all straight lines is: Obtain the minimum circumscribed rectangle of the welded steel structure region, draw a preset number of straight lines through the center point of the minimum circumscribed rectangle, and obtain the edge pixel points in the welded steel structure region; Record the two regions obtained by dividing any straight line as a first region and a second region, obtain the connecting line between any edge pixel point in the first region and the center point, and take the edge pixel point on the extension line of the connecting line in the second region as the symmetry point of the any edge pixel point; Calculate the difference amount of the number of edge pixel points between the first region and the second region, and record the distances between each edge pixel point in the first region, the symmetry point of each edge pixel point, and the center point as a first distance and a second distance respectively; calculate the sum of the differences between the first distance and the second distance of all edge pixel points in the first region; Calculate the inverse proportional mapping result of the product of the difference amount and the sum value; take the straight line with the maximum inverse proportional mapping result as the symmetry line; The process for obtaining the weld region and its characteristic length is: Obtain the connecting line of the two growth stop pixel points of each pixel point, and form the connecting line of all pixel points on the symmetry line into a weld region; and take the length with the maximum occurrence frequency in the lengths of the connecting lines of all pixel points on the symmetry line as the characteristic length.

2. A method of inspecting the quality of a welded steel structure of a construction work according to claim 1, characterized in that, The process for obtaining the growth stop pixel points is: For each pixel point, growth is performed in two directions perpendicular to the symmetry line, and each pixel point is taken as an initial growth pixel point, and the growth condition is that a gray difference value between a growth pixel point and its adjacent pixel point is calculated, and a ratio of the difference value to a gray value of the growth pixel point is less than or equal to a preset threshold value, and a last pixel point obtained by growth in each direction is taken as a growth stop pixel point.

3. A method of inspecting the quality of a welded steel structure of a construction work according to claim 1, characterized in that, The segmentation method of the local weld region is: Each segmentation pixel point is uniformly preset on the symmetry line, and a straight line passing through the segmentation pixel point and perpendicular to the symmetry line is used to segment the weld region into each local weld region.

4. The method for detecting the quality of a welded steel structure of a construction work according to Claim 1, wherein The acquisition process of the gray difference degree is: The maximum value of the gray difference between each pixel point and each preset adjacent pixel point in the direction of the symmetry line is counted, and the maximum value of the gray difference between each pixel point and each preset adjacent pixel point in the direction perpendicular to the symmetry line is counted. The gray difference degree is directly proportional to the maximum value and inversely proportional to the maximum value.

5. A method of inspecting the quality of a welded steel structure of a construction work according to claim 1, characterized in that, The expression of the texture feature value is: ; wherein, represents a texture feature value of the mth local weld seam area; represents the number of pixel points in the mth local weld seam area; represents the gray level difference degree of the nth pixel point in the mth local weld seam area; the minimum value in the deviation values of the gray level difference degrees between the nth pixel point in the mth local weld seam area and all the other pixel points is counted, represents the distance between the nth pixel point in the mth local weld seam area and the pixel point corresponding to the minimum value; σ represents a preset positive number; F represents the number of pixel points on the symmetry line in a single local weld seam area; represents the length of the connecting line between the fth pixel point on the symmetry line in the mth local weld seam area and the growth stop pixel point on the i side of the symmetry line; D represents the feature length; represents the absolute value operation.

6. A method of inspecting the quality of a welded steel structure of a construction work according to claim 1, characterized in that, The expression of the texture approximation degree is: ; wherein, represents the texture approximation degree between the local weld area A and the local weld area B; exp() represents the exponential function with the natural constant as the base number; , respectively represent the texture feature values of the local weld area A and the local weld area B; F represents the number of pixel points on the symmetry line in a single local weld area; , respectively represent the line lengths between the fth pixel point on the symmetry line in the local weld area A and the local weld area B and the growth stop pixel point on the ith side of the symmetry line; represents the absolute value operation.

7. A method of inspecting the quality of a welded steel structure of a construction work according to claim 1, characterized in that, The acquisition process of the welding quality evaluation value is: The mean value of the texture approximation degrees between each local weld region and all other local weld regions is calculated. The product of the texture feature value and the mean value is calculated. The discrete degree is mapped to a positive value, and the ratio of the product to the positive value is calculated. The welding quality evaluation value is a normalized value of the cumulative value of the ratios of all local weld regions.

8. A method of inspecting the quality of a welded steel structure of a construction work according to claim 1, characterized in that, The method for detecting the quality of the welded steel structure is: When the welding quality evaluation value is greater than or equal to a preset welding quality evaluation threshold value, it is determined that the quality of the welded steel structure is qualified, otherwise, it is determined that the quality is unqualified.

Images