Welding groove feature extraction method, system and welding robot

Abstract

The present invention provides a welding groove feature extraction method, system and welding robot, the method comprising: obtaining surface depth information data of the weld groove of a steel component and performing preprocessing to obtain the welding groove area, determining the position of the inflection point above the groove in any cross section in the welding groove area, judging and screening out standard welding groove cross sections, determining the position of the inflection point at the bottom of the groove in the cross section, obtaining the inflection point position information in each cross section, screening the inflection point position information in each cross section to obtain target inflection point position information in each cross section, judging the type of steel component, performing secondary processing on the target inflection point position information according to the steel component type, determining the start and end positions of the groove, and obtaining target groove feature information. The present invention adopts a data screening method to greatly reduce the data processing volume and improve the algorithm operation efficiency; at the same time, it adopts a multi-algorithm fusion to achieve accurate extraction of welding groove feature inflection points, thereby improving the accuracy and efficiency of welding groove feature extraction.

Description

Technical Field

[0001] The present invention relates to the technical field of visual inspection, and in particular to a welding groove inspection method, system and welding robot. Background Art

[0002] The steel structures in large buildings are mostly box-type steel columns or box-type beams, and their welding method is mostly butt welding. On construction sites, many large steel structures are still commonly welded manually. However, welding steel components in many high-rise buildings is dangerous and the welding conditions are quite harsh, which directly affects the welding efficiency and accuracy of welders. Compared with manual welding, welding robots have higher safety and work efficiency, and are more adaptable to the welding environment.

[0003] At present, more and more welding robots are used in standardized workshops, and the technology is becoming more and more mature. Welding robots need to complete the fully automatic intelligent welding of the grooves of steel structures. Among them, the automated detection of welding grooves is one of the important links. The detection accuracy and calculation efficiency of welding groove characteristics are also directly related to the welding quality and efficiency of welding robots.

[0004] However, when performing welding groove feature detection in the existing technology, the accuracy of welding groove feature extraction is often low due to inaccurate welding groove position positioning, and the algorithm for welding groove positioning is complex and requires large amounts of data to process, resulting in low overall calculation efficiency.

[0005] Therefore, how to improve the accuracy and efficiency of welding groove feature extraction is an urgent problem to be solved. Summary of the Invention

[0006] In view of this, it is necessary to provide a welding groove feature extraction method, system and welding robot to solve the problem of low accuracy and efficiency of welding groove feature extraction in the existing technology.

[0007] On the one hand, in order to achieve the above-mentioned object, the present invention provides a welding groove feature extraction method, comprising:

[0008] Obtain surface depth information data of the weld groove of steel components and perform preprocessing to obtain the weld groove area;

[0009] Determine the position of the inflection point above the groove in any cross section of the welding groove area;

[0010] The standard welding groove section is selected based on the position of the inflection point above the groove in the cross section;

[0011] Determine the inflection point position of the bottom of the groove in the cross section according to the screened standard welding groove cross section and the position of the inflection point above the groove in the cross section, and obtain the inflection point position information in each cross section;

[0012] Screening the inflection point position information within each cross section to obtain target inflection point position information within each cross section;

[0013] The type of the steel component is determined based on the target inflection point position information in the cross section, and the target inflection point position information is processed accordingly using a geometric method according to the type of the steel component to determine the starting and ending positions of the groove and obtain target groove feature information.

[0014] In some possible implementations, obtaining surface depth information data of a steel component weld groove and preprocessing the data to obtain a weld groove area includes:

[0015] Obtaining surface depth information data of the weld groove of the steel component, and calculating the Pearson correlation coefficient of the weld groove profile data of two adjacent sections in the weld groove surface depth information data, and determining the starting and ending positions of the weld groove using the Pearson correlation characteristics;

[0016] The welding groove area range is determined according to the starting and ending positions of the welding groove to obtain the welding groove area.

[0017] In some possible implementations, determining the position of the inflection point above the groove in any cross section in the welding groove area includes:

[0018] Based on the preset sliding window in the cross section, the absolute slope change of the data points on the left and right sides of the sliding window center point during each sliding is calculated, and the inflection point position above the groove in the cross section is determined according to the maximum value of the absolute slope change.

[0019] In some possible implementations, the step of selecting a standard welding groove section based on the position of the inflection point above the groove in the section includes:

[0020] The position of the inflection point above the groove in the cross section is obtained and a convex hull area judgment method is used to judge whether the welding groove cross section is standard, and the standard welding groove cross section is screened.

[0021] In some possible implementations, determining the position of the bottom inflection point of the groove in the cross section based on the screened standard welding groove cross section and the position of the upper inflection point of the groove in the cross section includes:

[0022] Obtaining the bottom contour of the groove in the standard welding groove cross section, selecting any two points on the bottom contour of the groove as preset inflection points, and calculating the sum of the distances between the points between adjacent inflection points and the straight lines on which the adjacent inflection points are located, based on the positions of the inflection points above the groove in the cross section;

[0023] Traverse all preset inflection point combinations, select the preset inflection point combination with the smallest sum of distances from the points between adjacent inflection points to the straight lines where the adjacent inflection points are located as the two bottom target inflection points, and determine the position of the bottom inflection point of the groove in the cross section.

[0024] In some possible implementations, filtering the inflection point position information within each cross section to obtain target inflection point position information within each cross section includes:

[0025] The RANSAC algorithm is used to filter the inflection point position information in each section, and a first-order linear model of the position of the inflection point in the section, the depth value and the position value of the groove direction in the inflection point position information is established. Based on the first-order linear model, the relatively discrete abnormal data are removed to obtain the target inflection point position information in each section.

[0026] In some possible implementations, the determining the type of the steel member based on the inflection point position information of the target within the cross section, and performing corresponding processing on the target inflection point position information using a geometric method according to the steel member type to determine the start and end positions of the groove, thereby obtaining target groove feature information, includes:

[0027] Obtaining inflection point position information of the target within the cross section, calculating a first-order linear equation of the straight line portion at the middle position of the weld groove using a least squares method, and calculating the sum of the distances from the inflection points on both sides of the straight line where the first-order linear equation is located to the straight line where the first-order linear equation is located;

[0028] If the sum of the distances from the inflection points on both sides of the line where the first-order linear equation is located to the line where the first-order linear equation is located is greater than a preset threshold, then the welding surface of the steel member is not a plane and is a rounded steel member; otherwise, the welding surface of the steel member is a plane and is a right-angled steel member;

[0029] Performing secondary processing on the groove feature information of the rounded corners on two adjacent sides of the rounded corner steel member using a geometric method to determine the specific position information of the rounded corners on both sides of the groove;

[0030] The starting and ending positions of the groove of the rounded-angle steel member or the right-angle steel member are determined to obtain target groove feature information.

[0031] In some possible implementations, the target groove feature information includes:

[0032] The position of each inflection point of the welding groove, the depth value corresponding to each inflection point of the welding groove, and the position of each welding groove section.

[0033] On the other hand, the present invention also provides a welding groove feature extraction system, which includes a memory, a processor, and a computer program stored in the memory and runnable on the processor. When the processor executes the program, it implements a welding groove feature extraction method described in the above implementation method.

[0034] On the other hand, the present invention further provides a welding robot, comprising a welding groove feature extraction system described in the above implementation manner.

[0035] The beneficial effect of adopting the above embodiment is: the welding groove feature extraction method provided by the present invention determines the position information of each inflection point in the welding groove area through the surface depth information data of the weld groove of the steel component, and distinguishes and processes different types of steel components in combination with the inflection point characteristics, and then extracts the groove feature information separately. At the same time, data screening and exception processing operations are performed multiple times during the data processing process, while retaining the feature information and removing redundant information and abnormal information. On the one hand, the calculation amount of welding groove feature extraction is reduced, and the efficiency of welding groove feature extraction is improved. On the other hand, the processing of abnormal information and the distinction between different types of steel components improve the accuracy of welding groove feature extraction. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0037] Figure 1 A schematic flow chart of an embodiment of a welding groove detection method provided by the present invention;

[0038] Figure 2 A schematic diagram of an embodiment of scanning a welding groove of a steel member provided by the present invention;

[0039] Figure 3 A schematic diagram of an embodiment of the scanning direction of the welding groove of a steel member provided by the present invention;

[0040] Figure 4 A schematic diagram of an embodiment of a method for determining the inflection point of a welding groove of a steel member provided by the present invention;

[0041] Figure 5 A schematic flow chart of an embodiment of step S106 provided by the present invention;

[0042] Figure 6 This is a structural diagram of an embodiment of a right-angle box-shaped steel column and a rounded-corner box-shaped steel column provided by the present invention;

[0043] Figure 7 A schematic structural diagram of an embodiment of the feature information of a rounded box-shaped steel provided by the present invention;

[0044] Figure 8 This is a schematic structural diagram of an embodiment of the welding groove feature extraction system provided by the present invention. DETAILED DESCRIPTION

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0046] It should be understood that the schematic drawings are not drawn to scale. The flowcharts used in the present invention illustrate operations implemented according to some embodiments of the present invention. It should be understood that the operations in the flowcharts may be implemented out of sequence, and steps that do not have a logical contextual relationship may be reversed or performed simultaneously. In addition, those skilled in the art, guided by the present disclosure, may add one or more additional operations to the flowcharts or remove one or more operations from the flowcharts.

[0047] Some of the blocks shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor systems and / or microcontroller systems.

[0048] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0049] Embodiments of the present invention provide a welding groove detection method, system, and welding robot, which are described below respectively.

[0050] like Figure 1 As shown, Figure 1 A schematic flow chart of an embodiment of a welding groove detection method provided by the present invention includes:

[0051] S101, obtaining surface depth information data of the weld groove of the steel component and performing preprocessing to obtain the welding groove area;

[0052] S102, determining the position of the inflection point above the groove in any cross section of the welding groove area;

[0053] S103, judging and screening a standard welding groove section according to the position of the inflection point above the groove in the cross section;

[0054] S104, determining the position of the inflection point at the bottom of the groove in the cross section according to the screened standard welding groove cross section and the position of the inflection point above the groove in the cross section, and obtaining the inflection point position information in each cross section;

[0055] S105, screening the inflection point position information in each cross section to obtain target inflection point position information in each cross section;

[0056] S106. Determine the type of the steel member based on the target inflection point position information in the cross section, and perform corresponding processing on the target inflection point position information using a geometric method according to the type of the steel member to determine the start and end positions of the groove, and obtain target groove feature information.

[0057] The embodiment of the present invention determines the position information of each inflection point in the welding groove area through the surface depth information data of the weld groove of the steel component, and distinguishes and processes different types of steel components in combination with the inflection point characteristics, and then extracts the groove feature information respectively. At the same time, data screening and exception processing operations are performed multiple times during the data processing process. While retaining the feature information, redundant information and abnormal information are removed. On the one hand, the calculation amount of welding groove feature extraction is reduced and the efficiency of welding groove feature extraction is improved. On the other hand, the processing of abnormal information and the distinction between different types of steel components improve the accuracy of welding groove feature extraction.

[0058] In some embodiments of the present invention, step S101 includes:

[0059] Obtaining surface depth information data of the weld groove of the steel component, and calculating the Pearson correlation coefficient of the weld groove profile data of two adjacent sections in the weld groove surface depth information data, and determining the starting and ending positions of the weld groove using the Pearson correlation characteristics;

[0060] The welding groove area range is determined according to the starting and ending positions of the welding groove to obtain the welding groove area.

[0061] In the specific embodiment of the present invention, please refer to Figure 2 , Figure 2This is a schematic diagram of an embodiment of scanning the welding groove of a steel component provided by the present invention. A welding robot uses a line structured light vision sensor to scan the surface to be welded of a large building steel component. The scanning result can be generated as 1200*1200 (X-axis and Y-axis directions) point cloud data, where each row of data represents a cross-section of the scanned groove and contains the contour shape information of the welding groove.

[0062] Furthermore, based on the surface depth information data of the weld groove, the Pearson correlation coefficient is used to calculate and filter the weld groove area data, and the abnormal points scanned in the weld groove area data are processed to filter the weld groove area from the original data. The specific process is as follows:

[0063] The Pearson correlation coefficient is used to measure the degree of linear correlation between two variables, and its value is between -1 and 1. The intuitive expression of this linear correlation is whether Y increases or decreases at the same time as X increases; when the two are distributed on a straight line, the Pearson correlation coefficient is equal to 1 or -1; there is no linear relationship between the two variables, and the Pearson correlation coefficient is 0. Extract the contour data of the welding groove of two adjacent sections, and assume that the contour data of the welding groove of two adjacent sections are X and Y (X, Y generally contain 1200 contour data points), and perform similarity matching on them according to the formula. Scan the welding groove from left to right and calculate the correlation coefficient of all adjacent welding groove contours. The formula for the Pearson correlation coefficient is as follows:

[0064]

[0065] Where X and Y represent the weld groove profile data of two adjacent sections in the scanning depth information, i represents the i-th data in X or Y, and n represents the total number of data points in the selected column (usually 1200). In the above formula, μ X and μ Y are the means of X and Y, σ X and σ Y are the standard deviations of X and Y, respectively, where:

[0066]

[0067]

[0068] See also Figure 3 , Figure 3This is a schematic diagram of an embodiment of the scanning direction for a steel component weld groove provided by the present invention. When the line structured light scanning beam is at position A, the line structured light fringes before and after position A change slowly, and the line structured light scanning profile is a straight line. When moving along the scanning direction, at the weld groove starting point position B, the line structured light scanning profile differs from the line structured light scanning profile at the previous moment, resulting in a smaller Pearson correlation coefficient corr. Continuing to position C, due to the structural characteristics of the weld groove, the line structured light scanning profile near position C differs slightly from the line structured light scanning profile at the previous moment, resulting in a larger Pearson correlation coefficient corr. When reaching the weld groove end point D, the line structured light scanning profile differs from the line structured light scanning profile at the previous moment, resulting in a smaller Pearson correlation coefficient corr. When leaving the weld groove, position E is a flat area, the line structured light scanning profile is a straight line, and the Pearson correlation coefficient corr is large. From the above analysis, it can be seen that the area between the two locations with smaller Pearson correlation coefficients corr is the weld groove area.

[0069] It should be noted that the laser emitted by the line structured light vision sensor may produce abnormal scanning results due to problems such as scanning angle. This is manifested as a value of -10000 in the point cloud data. To facilitate subsequent processing, this abnormal value in the area where the welding groove is located is assigned a value of -200.

[0070] The embodiment of the present invention provides a method for screening the welding groove area by using the Pearson correlation coefficient and processing the outliers, which reduces the amount of data calculation on the one hand and reduces the influence of the outliers on the welding groove feature extraction on the other hand.

[0071] Furthermore, in some embodiments of the present invention, step S102 includes:

[0072] Based on the preset sliding window in the cross section, the absolute slope change of the data points on the left and right sides of the sliding window center point during each sliding is calculated, and the inflection point position above the groove in the cross section is determined according to the maximum value of the absolute slope change.

[0073] In the specific embodiment of the present invention, please refer to Figure 4 , Figure 4 This is a schematic diagram of an embodiment of the method for determining the inflection point of a steel member welding groove provided by the present invention. The position of the inflection point above the groove in the cross section is calculated and determined based on the welding groove area data: first, the position of the inflection point point1 above the groove on the left is obtained, a sliding window with a length of (2a+1) is set, and the average slope k of the data points on the left and right sides of the window center point (i, z(i)) is calculated. left and k right , and find the absolute change in slope Δk:

[0074]

[0075] Δk=|k left -k right |

[0076] Where a represents the number of data points to the left (right) of the window center, (2a+1) represents the window length, z(i+1),...,z(i+a / 2-1),z(i+a / 2+1),...,z(i+a) represent the points to the right of the window center, and z(ia),...,z(ia / 2-1),z(ia / 2+1),...,z(i-1) represent the points to the left of the window center. The window slides from left to right within the groove section, and Δk is calculated with each slide. The calculation result and the corresponding position are recorded in a two-dimensional array. The Δk corresponding to inflection point point1 is the maximum value within the range, so the location of inflection point point1 can be determined by searching for the position of Δkmax. After determining the location of inflection point point1 above the groove, a sliding window of length (2a+1) is set, starting from the Sth point to the right of point1, and sliding from right to left. The location of inflection point point2 is determined in the same manner as described above.

[0077] Furthermore, in some embodiments of the present invention, step S103 includes:

[0078] The position of the inflection point above the groove in the cross section is obtained and a convex hull area judgment method is used to judge whether the welding groove cross section is standard, and the standard welding groove cross section is screened.

[0079] In a specific embodiment of the present invention, the trapz() function is first used to calculate the actual area s1 of the groove between point1 and point2. The trapz() function uses the trapezoidal method to perform numerical integration to obtain the actual area of ​​the groove. The convex hull function convhull() is then used to calculate the convex hull area s2 of the groove.

[0080] This method uses Graham Scan as a convex hull calculation method. The algorithm works as follows: First, a starting point and a next starting point are selected, typically the leftmost or rightmost point in the point set. Then, the traversal is performed in a specific order. If the convexity of the "shell" formed by the newly added point and the previously selected point remains unchanged, the traversal continues. Otherwise, the last found point is deleted and the convexity is compared again. The traversal continues until the convexity remains unchanged. Finally, the upper and lower "shells" are scanned separately and merged to obtain the convex hull of the weld groove data. The area s2 of the convex hull is then calculated, ultimately yielding the area ratio s = s2 / s1.

[0081] According to the actual situation of the welding groove at the welding site, if s>1.1, the groove in the section can be considered as a non-standard groove; the closer s is to 1, the more standard the groove shape is. A mark is set for the groove situation in each section. The following steps are all processed for the standard groove section type.

[0082] Furthermore, in some embodiments of the present invention, step S104 includes:

[0083] Obtain the bottom contour of the groove in the standard welding groove section, select any two points on the bottom contour of the groove as preset inflection points, combine the position of the inflection point above the groove in the section, calculate the sum of the distances between the points between adjacent inflection points and the straight lines where the adjacent inflection points are located, traverse all preset inflection point combinations, select the preset inflection point combination with the smallest sum of the distances as the two target inflection points at the bottom, and determine the position of the inflection point at the bottom of the groove in the section.

[0084] In a specific embodiment of the present invention, the characteristic inflection point at the bottom of the groove is determined by using a DP thinning algorithm to thin the data, setting a threshold so that the number of data points within the groove is less than 20. The DP thinning algorithm can be used to compress a large number of redundant graphic data points to extract the necessary data points.

[0085] The DP algorithm process is as follows:

[0086] Connect the first and last points A and B of the curve.

[0087] Calculate the distances from all points on the curve to the curves where points A and B are located.

[0088] Calculate the maximum distance D. If D is less than the threshold, remove all points on the curve except A and B. If D is greater than the threshold, split the curve into two segments with the maximum distance.

[0089] Repeat the above steps for all curve segments until all D are smaller than the threshold, completing the thinning.

[0090] Find the bottom inflection points point3 and point4:

[0091] Assume the equation of the straight line where the three sides determined by the four inflection points of the groove are located:

[0092]

[0093] The three straight line equations above are the straight lines corresponding to the inflection points point1 and point2, point2 and point3, and point3 and point4. Calculate the sum of the distances from the points between adjacent inflection points to the corresponding straight line equations:

[0094]

[0095] Traverse all combinations of point3 and point4, calculate the corresponding distance sum e, and select the two points that make e the smallest as the inflection points of the bottom of the welding groove, and obtain the inflection point information of the bottom of the welding groove.

[0096] Furthermore, in some embodiments of the present invention, step S105 includes:

[0097] The RANSAC algorithm is used to filter the inflection point position information in each section, and a first-order linear model of the position of the inflection point in the section, the depth value and the position value of the groove direction in the inflection point position information is established. Based on the first-order linear model, the relatively discrete abnormal data are removed to obtain the target inflection point position information in each section.

[0098] In a specific embodiment of the present invention, in order to avoid abnormal inflection points that deviate from a normal range in the extracted feature data, the RANSAC algorithm is used to screen the extracted groove feature inflection point data.

[0099] Based on the above algorithm, a first-order linear model is established for the cross-sectional location of the inflection point, the depth value, and the groove direction (i.e., the Y and X directions, and the Y and Z directions) in the characteristic data. A reasonable threshold is set to remove abnormal data that deviates significantly from the normal range. The thresholds for the Y and Z direction models are appropriately increased to maximize the preservation of the extracted features of the fillet position on both sides of the groove of the rounded box-shaped steel member.

[0100] The algorithm can iteratively estimate the parameters of the mathematical model from a set of observation data containing "outliers". Its basic assumptions are: the data consists of "inliers"; "outliers" are data that cannot adapt to the model; and data other than these are noise.

[0101] The embodiment of the present invention further screens the welding groove feature data through the DP thinning algorithm and the RANSAC algorithm, while retaining the feature data, eliminating redundant and abnormal data, reducing the amount of calculation, and further improving the efficiency of groove feature inflection point extraction.

[0102] Furthermore, in some embodiments of the present invention, Figure 5 As shown, Figure 5 This is a flow chart of an embodiment of step S106 provided by the present invention. Step S106 includes:

[0103] S501, obtaining inflection point position information of the target in the cross section, calculating the first-order linear equation of the straight line portion at the middle position of the welding groove using the least squares method, and calculating the sum of the distances from the inflection points on both sides of the straight line where the first-order linear equation is located to the straight line where the first-order linear equation is located;

[0104] S502: If the sum of the distances from the inflection points on both sides of the line containing the first-order linear equation to the line containing the first-order linear equation is greater than a preset threshold, then the weld surface of the steel member is not a plane and is a rounded-angle steel member; otherwise, the weld surface of the steel member is a plane and is a right-angle steel member;

[0105] S503, performing secondary processing on the groove feature information of the rounded corners on two adjacent sides of the rounded corner steel member using a geometric method to determine specific position information of the rounded corners on both sides of the groove;

[0106] S504: Determine the starting and ending positions of the groove of the rounded-angle steel member or the right-angle steel member to obtain target groove feature information.

[0107] In a specific embodiment of the present invention, the type of the scanned box-shaped steel member is determined: Figure 6 , Figure 6 This is a structural schematic diagram of an embodiment of the right-angle box-shaped steel column and the rounded-corner box-shaped steel column provided by the present invention, wherein the box-shaped steel components are generally divided into right-angle box-shaped steel columns and rounded-corner box-shaped steel columns.

[0108] The least squares method is used to calculate the first-order linear equation of the characteristic inflection point of the straight groove in the middle of the welding groove, that is, the linear equation of the characteristic inflection point Y-axis and Z-axis:

[0109] Ay+Bz+C=0

[0110] Calculate the sum of the distances e′ from the inflection points on both sides of the extracted groove feature (10 data points) to the above first-order straight line equation:

[0111]

[0112] In the above formula, n is the total number of sections of characteristic data extracted from the welding groove, (y i ,z i ) represents the Y-axis direction data and Z-axis direction depth value corresponding to the left groove feature inflection point, (y j ,z j ) represents the Y-axis direction data and Z-axis direction depth value corresponding to the groove feature inflection point on the right side, e left represents the sum of the distances from the 10 groove feature inflection points on the left to the above straight line, e right Represents the sum of the distances from the 10 groove feature inflection points on the right to the above straight line.

[0113] To more clearly demonstrate the characteristics of rounded steel members, see Figure 7 , Figure 7 This is a schematic diagram of an embodiment of the characteristic information of rounded box steel provided by the present invention, which is composed of Figure 7It can be seen that if e′ is greater than the given threshold T, it can be considered that the welding surface of the box-section steel member is not a plane and there is a rounded corner, that is, the box-section steel member is a rounded steel member; if e′ is less than or equal to T, it can be considered that the welding surface of the box-section steel member is close to a plane and there is no rounded corner, that is, the box-section steel member is a right-angle steel member.

[0114] For right-angle steel components, combined with the original point cloud data, the starting and ending positions (Y-axis direction) of the extracted groove features are traversed within the range of [-20, 20] to accurately determine the starting and ending positions of the groove, and obtain the welding groove feature extraction results of the right-angle steel components.

[0115] If it is a rounded box-shaped steel member, find the specific position information of the rounded corners on both sides of the groove:

[0116] Starting from the determined starting and ending positions of the welding groove area (i.e. the left starting position and the right ending position of the groove), and within the determined groove position area (between point and point2 in the X-axis direction), traverse the average slope of the 10 adjacent points on the inner side of the point at the fillet position from the outside to the inside:

[0117]

[0118] In the above formula, z(i+1), z(i+2), ..., z(i+10) represent the depth values ​​of the points at the 10 rounded corners on the left, z(j-10), ..., z(j-2), z(j-1) represent the depth values ​​of the points at the 10 rounded corners on the right, and k left represents the average slope determined by 10 points on the left corner, k right It represents the average slope determined by the 10 points on the right side of the fillet position. If the average slope of a point is close to the slope of the determined straight line equation, it can be considered that the point is located at the intersection of the fillet part of the box-section steel member and the middle straight line part, and the range of the fillet part can be determined. Then, the positions where the slopes are 1 and -1 are searched respectively within the fillet range on both sides, thereby determining the position information of the fillet parts on both sides of the groove.

[0119] Furthermore, the groove features of the rounded corners of the box-shaped steel components are processed twice:

[0120] According to the determined left and right round corner positions, the polyfit() function is used to obtain the quadratic fitting curve equation of the round corner profile, which can be regarded as the fitting curve equation of the Y-axis and Z-axis values ​​of the two inflection points point1 and point2 above the groove.

[0121] Based on the groove characteristics at the middle position of the groove, the average depth of the groove is extracted. The middle position of the groove is selected as (m, m+1, m+2,..., m+l), d13 and d 24 Respectively represent the groove depth of the right-angled side and the hypotenuse in the Z-axis direction, z1(i), z3(i) represent the depth values ​​of the inflection points point1 and point3 on the right-angled side in the i-th groove section, z2(j), z4(j) represent the depth values ​​of the inflection points point2 and point4 on the hypotenuse in the j-th groove section:

[0122]

[0123] According to the obtained groove depth and the determined fitting curve, move down d 13 and d 24 The fitting curve equations of the Y-axis and Z-axis values ​​on both sides of the bottom of the groove (both sides of the bottom of the groove where point3 and point4 are located) can be obtained.

[0124] According to the groove features extracted from the straight groove in the middle position, the first-order linear equations in the Y-axis and X-axis directions of the four groove inflection points are obtained using the least squares method, that is, the straight line fitting equations of the four sides of the middle position of the welding groove.

[0125] The fitting line equation determined above is used to determine the position information (i.e., Y and X) and depth information (i.e., Y and Z) of the four inflection points at the fillet positions on the left and right sides of the groove (only for groove sections where the groove features are not extracted) to obtain the groove features at the fillet positions. Considering the splicing problem of the welds on two adjacent sides of the box-section steel component, the positions where the slopes of the fillet contour curve are 1 and -1 are taken as the edge positions of the fillet groove features. Finally, the groove feature data of the fillet positions on both sides and the groove in the middle are spliced ​​together to obtain the welding groove feature extraction results of the fillet box-section steel component.

[0126] The embodiment of the present invention distinguishes different types of steel components by calculating and analyzing the characteristics of the inflection points of the welding groove characteristics of the steel components, thereby further improving the accuracy of welding groove feature extraction.

[0127] like Figure 8 As shown, the present invention also provides a welding groove feature extraction system 800. The system 800 includes a processor 801, a memory 802 and a display 803. Figure 8 Only some of the components of system 800 are shown, but it should be understood that implementing all of the shown components is not a requirement, and more or fewer components may alternatively be implemented.

[0128] In some embodiments, the processor 801 may be a central processing unit (CPU), a microprocessor, or other data processing chip, configured to execute program codes or process data stored in the memory 802, such as the welding groove feature extraction method of the present invention.

[0129] In some embodiments, processor 801 may be a single server or a server group. The server group may be centralized or distributed. In some embodiments, processor 801 may be local or remote. In some embodiments, processor 801 may be implemented on a cloud platform. In one embodiment, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an on-premises cloud, a multi-cloud, or any combination thereof.

[0130] In some embodiments, the memory 802 may be an internal storage unit of the system 800, such as a hard disk or memory of the system 800. In other embodiments, the memory 802 may also be an external storage device of the system 800, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc., equipped on the system 800.

[0131] Furthermore, the memory 802 may include both an internal storage unit of the system 800 and an external storage device. The memory 802 is used to store application software installed in the system 800 and various data.

[0132] In some embodiments, display 803 can be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 803 is used to display information on system 800 and to display a visual user interface. Components 801-803 of system 800 communicate with each other via a system bus.

[0133] In one embodiment, when the processor 801 executes the welding groove feature extraction program in the memory 802, the following steps may be implemented:

[0134] Obtain surface depth information data of the weld groove of steel components and perform preprocessing to obtain the weld groove area;

[0135] Determine the position of the inflection point above the groove in any cross section of the welding groove area;

[0136] The standard welding groove section is selected based on the position of the inflection point above the groove in the cross section;

[0137] Determine the inflection point position of the bottom of the groove in the cross section according to the screened standard welding groove cross section and the position of the inflection point above the groove in the cross section, and obtain the inflection point position information in each cross section;

[0138] Screening the inflection point position information within each cross section to obtain target inflection point position information within each cross section;

[0139] The type of the steel component is determined based on the target inflection point position information in the cross section, and the target inflection point position information is processed accordingly using a geometric method according to the type of the steel component to determine the starting and ending positions of the groove and obtain target groove feature information.

[0140] It should be understood that, when the processor 801 executes the welding groove feature extraction program in the memory 802 , in addition to the above functions, it can also implement other functions. For details, please refer to the description of the corresponding method embodiment above.

[0141] The welding groove feature extraction system 800 provided in the above embodiment can implement the technical solution described in the above welding groove feature extraction method embodiment. The specific implementation principles of the above components can be found in the corresponding contents in the above welding groove feature extraction system embodiment, which will not be repeated here.

[0142] Correspondingly, an embodiment of the present application also provides a welding robot, including the welding groove feature extraction system provided in the above embodiment.

[0143] In a specific embodiment, the welding robot may be a robot including a welding groove feature extraction system, which is not limited here. A welding robot using the welding groove feature extraction system can realize intelligent, efficient and high-quality welding operations at the construction site.

[0144] Those skilled in the art will appreciate that all or part of the process steps of the above-described embodiments can be implemented by instructing related hardware (such as a processor, a controller, etc.) through a computer program, and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a magnetic disk, an optical disk, a read-only memory, or a random access memory.

[0145] The above is a detailed introduction to the welding groove feature extraction method, system and welding robot provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for those skilled in the art, according to the ideas of the present invention, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A welding groove feature extraction method, characterized in that: include: Obtaining surface depth information data of the weld groove of the steel component, and calculating the Pearson correlation coefficient of the weld groove profile data of two adjacent sections in the weld groove surface depth information data, and determining the starting and ending positions of the weld groove using the Pearson correlation characteristics; Determine the welding groove area range according to the welding groove starting and ending positions to obtain the welding groove area; Determine the position of the inflection point above the groove in any cross section of the welding groove area; The standard welding groove section is selected based on the position of the inflection point above the groove in the cross section; Determine the inflection point position of the bottom of the groove in the cross section according to the screened standard welding groove cross section and the position of the inflection point above the groove in the cross section, and obtain the inflection point position information in each cross section; Obtain the bottom contour of the groove in the standard welding groove cross section, select any two points on the bottom contour of the groove as preset inflection points, and calculate the sum of the distances between the points between adjacent inflection points and the straight lines where the adjacent inflection points are located, based on the positions of the inflection points above the groove in the cross section; Traversing all preset inflection point combinations, selecting the preset inflection point combination with the smallest sum of distances from points between adjacent inflection points to the straight lines where the adjacent inflection points are located as the two bottom target inflection points, and determining the position of the bottom inflection point of the groove in the cross section; Obtaining inflection point position information of the target within the cross section, calculating a first-order linear equation of the straight line portion at the middle position of the weld groove using a least squares method, and calculating the sum of the distances from the inflection points on both sides of the straight line where the first-order linear equation is located to the straight line where the first-order linear equation is located; If the sum of the distances from the inflection points on both sides of the line where the first-order linear equation is located to the line where the first-order linear equation is located is greater than a preset threshold, then the welding surface of the steel member is not a plane and is a rounded steel member; otherwise, the welding surface of the steel member is a plane and is a right-angled steel member; Performing secondary processing on the groove feature information of the rounded corners on two adjacent sides of the rounded corner steel member using a geometric method to determine the specific position information of the rounded corners on both sides of the groove; The starting and ending positions of the groove of the rounded-angle steel member or the right-angle steel member are determined to obtain target groove feature information.

2. A welding groove feature extraction method according to claim 1, characterized in that: Determining the position of the inflection point above the groove in any cross section of the welding groove area includes: Based on the preset sliding window in the cross section, the absolute slope change of the data points on the left and right sides of the sliding window center point during each sliding is calculated, and the inflection point position above the groove in the cross section is determined according to the maximum value of the absolute slope change.

3. A welding groove feature extraction method according to claim 1, characterized in that: The step of judging and selecting a standard welding groove section based on the position of the inflection point above the groove in the section includes: The position of the inflection point above the groove in the cross section is obtained and a convex hull area judgment method is used to judge whether the welding groove cross section is standard, and a standard welding groove cross section is obtained by screening.

4. A welding groove feature extraction method according to claim 1, characterized in that: The inflection point position information in each cross section is screened to obtain target inflection point position information in each cross section, including: The RANSAC algorithm is used to filter the inflection point position information in each section, and a first-order linear model of the position of the inflection point in the section, the depth value and the position value of the groove direction in the inflection point position information is established. Based on the first-order linear model, the relatively discrete abnormal data are removed to obtain the target inflection point position information in each section.

5. A welding groove feature extraction method according to claim 1, characterized in that: The target groove feature information includes: The position of each inflection point of the welding groove, the depth value corresponding to each inflection point of the welding groove, and the position of each welding groove section.

6. A welding groove feature extraction system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, a welding groove feature extraction method according to any one of claims 1 to 5 is implemented.

7. A welding robot, characterized in that: It includes a welding groove feature extraction system as described in claim 6.

Images