A method and system for feature extraction and dimensional measurement of saddle-shaped weld grooves in thick plates

Abstract

This application belongs to the field of visual recognition technology, specifically disclosing a method and system for feature extraction and size measurement of saddle-shaped weld grooves in thick plates. It includes: based on multi-scale geometric feature abrupt change detection, segmenting the current frame's weld groove 3D point cloud into four regions: branch pipe surface, groove sidewall, root pass weld surface, and main pipe surface, obtaining four 3D point clouds; based on the 3D point clouds of each region, determining the intersection points of the branch pipe surface and groove sidewall, the groove sidewall and root pass weld surface, the root pass weld surface and main pipe surface, the intersection points of the branch pipe surface and main pipe surface, and the virtual intersection points of the extension lines of the groove sidewall and main pipe surface, collectively serving as weld groove feature points; constructing vectors between feature points, and calculating the weld groove geometric dimensions through vector operations. This application, based on multi-scale geometric feature abrupt change detection and adaptive point cloud segmentation, does not require a preset fitting curve slope threshold and can adapt to different angles between the branch pipe and main pipe at pipe-to-pipe intersection nodes.

Description

Technical Field

[0001] This application belongs to the field of visual recognition technology, and more specifically, relates to a method and system for extracting features and measuring dimensions of saddle-shaped weld grooves in thick plates. Background Technology

[0002] Automated welding requires pre-weld bevel feature recognition. This recognition identifies the weld center point, and the welding robot welds along this center point. Existing weld bevel recognition methods primarily target V-grooves for straight and fillet welds. These methods use the weld center point as the feature extraction point and calculate the bevel opening width as the bevel geometry. For right-angle and straight welds, the bevel appearance and geometry remain fixed along a single spatial direction. V-grooves exhibit distinct geometric features and symmetry; their components can be fitted with straight lines, and feature points are determined by the intersection of the fitted lines or directly by the location of point cloud difference abrupt changes.

[0003] Ding's team proposed an online shape matching algorithm that determines U-shaped / V-shaped bevel features through correlation coefficient thresholds, achieving an adaptability of 89%. Geng et al. used an improved RANSAC multi-plane fitting algorithm to accurately obtain the intersection line positions of intersecting planes in point cloud models for three different medium-thick plate structural component models (three-plane butt welding, H-beam butt welding, and square tube butt welding), and extracted weld seams by combining geometric features. Based on the geometric features of the three models, they proposed specific methods for extracting weld seam geometric features. Zhang et al. proposed a three-line laser segmented fitting strategy, combined with Freeman chain codes to eliminate outliers, achieving stable feature extraction under dynamic working conditions.

[0004] However, the aforementioned studies mostly focus on planar straight welds and fillet welds, relying on fixed geometric features. In contrast, the groove of a spatially intersecting, multi-layered, multi-pass variable cross-section saddle-shaped weld exhibits a spatially varying curve, with complex geometry and spatial diversity, and variable groove opening directions. These methods may need to readjust the fitting equations and correlation coefficient thresholds when dealing with the changing geometric features of saddle-shaped weld grooves. Furthermore, current groove feature recognition primarily relies on line laser scanning and image recognition. The scanning and imaging angles must remain perpendicular to the weld at all times. Therefore, the recognition process relies solely on two-dimensional images or two-dimensional point cloud computing. Existing methods cannot adapt to the feature point recognition and geometric feature calculation of asymmetric variable cross-section saddle-shaped grooves with varying three-dimensional spatial positions. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the purpose of this application is to provide a method and system for extracting features and measuring dimensions of saddle-shaped weld grooves in thick plates, aiming to solve the problem that existing weld groove feature extraction point identification methods cannot be applied to saddle-shaped weld grooves.

[0006] To achieve the above objectives, in a first aspect, this application provides a method for extracting features and measuring dimensions of a saddle-shaped weld groove in a thick plate, comprising:

[0007] Obtain the 3D point cloud of the saddle-shaped weld groove in a thick plate;

[0008] Based on multi-scale geometric feature mutation detection, the current frame weld groove three-dimensional point cloud is segmented into branch pipe surface, groove sidewall, root pass surface and main pipe surface, resulting in four regions of three-dimensional point cloud;

[0009] Based on the three-dimensional point cloud of each region, the intersection points of the branch pipe face and the bevel sidewall, the intersection points of the bevel sidewall and the root pass weld surface, the intersection points of the root pass weld surface and the main pipe face, the intersection points of the branch pipe face and the main pipe face, and the virtual intersection points of the bevel sidewall and the extension line of the main pipe face are determined and used together as the weld bevel feature points.

[0010] Construct vectors between feature points, and calculate the geometric dimensions of the weld groove through vector operations.

[0011] Preferably, the multi-scale geometric feature abrupt change detection divides the current frame weld groove 3D point cloud into four regions: branch pipe surface, groove sidewall, root pass weld surface, and main pipe surface, as detailed below:

[0012] Principal component analysis is performed on the neighborhood point set of each point in the current frame's 3D point cloud to obtain the eigenvalues ​​and corresponding eigenvectors of the covariance matrix in the principal component analysis.

[0013] The eigenvector corresponding to the smallest eigenvalue is taken as a point. Normal vector at the weld bevel ;

[0014] Calculate the proportion of the smallest eigenvalue to the sum of all eigenvalues, and use this proportion as the point. Curvature at the weld bevel ;

[0015] Calculate the change in the angle between the normal vectors of adjacent points ,in, For vector dot product, To take the absolute value;

[0016] Calculate curvature difference ;

[0017] Constructing a comprehensive feature function , These are the weighting coefficients, ;

[0018] like Judgment point This is the local maximum value point at the weld bevel.

[0019] Sort all local maximum points in descending order of comprehensive characteristic value, and retain the three largest ones as the dividing points;

[0020] The three detected segmentation points are indexed as segmentation point locations, and the three-dimensional point cloud of the weld groove is divided into four three-dimensional point clouds.

[0021] Preferably, the process of determining the characteristic points of the weld groove is as follows:

[0022] (1) The intersection of the branch pipe face and the bevel sidewall, the intersection of the bevel sidewall and the root pass weld surface, and the intersection of the root pass weld surface and the main pipe face are respectively represented as the length of the fitting window correction.

[0023] (2) Based on the intersection of the three-dimensional point cloud of each region and the above-mentioned intersection with the fitted window correction length, determine the intersection of the pipe surface and the bevel sidewall, the intersection of the bevel sidewall and the root pass weld surface, the intersection of the root pass weld surface and the main pipe surface, the intersection of the branch pipe surface and the main pipe surface, and the virtual intersection of the bevel sidewall and the extension line of the main pipe surface.

[0024] (3) Calculate the minimum distance deviation between the current frame's 3D point cloud and the four intersection points determined in this round;

[0025] (4) If the minimum distance deviation is greater than the preset value, adjust at least one fitting window correction length value and return to step (2); otherwise, take all current intersection points as feature points.

[0026] It should be noted that this application judges the validity of the solved values ​​by the distance deviation between the intersection point and the real point cloud, and corrects the position of the feature points to achieve accurate calculation of the geometric dimensions of the bevel and precise positioning of the feature points.

[0027] Preferably, adjusting at least one fitting window correction length value specifically involves:

[0028] If the minimum distance deviation corresponds to the bevel sidewall, then fine-tune the fitting window correction length of the intersection of the bevel sidewall and the fitting window correction length of the intersection of the bevel sidewall and the root pass surface to make the bevel surface point cloud more coplanar.

[0029] If the distance between the intersection of the bevel sidewall and the root pass surface and the intersection of the root pass surface and the main pipe surface is greater than a preset threshold, then the fitting window correction length of the intersection of the bevel sidewall and the root pass surface and the fitting window correction length of the intersection of the root pass surface and the main pipe surface are finely adjusted so that the distance between the intersection of the bevel sidewall and the root pass surface and the intersection of the root pass surface and the main pipe surface does not exceed the theoretical value range of the saddle-shaped weld gap weld width.

[0030] Preferably, the weld groove geometry includes: the width of the root pass weld surface, the groove depth, the groove angle, the groove width, the groove cross-sectional area, the angle between the groove sidewall and the root pass weld surface, and the angle between the main pipe surface and the root pass weld surface.

[0031] It should be noted that this application, based on accurately identified feature points, further calculates the geometric dimensions of the weld bevel, including the weld root pass width, bevel depth, bevel angle, bevel cross-sectional area, and the angle between the root pass surface and the main and branch pipes, as well as the welding torch posture. The measurement results cover the key data required for welding process parameter optimization and welding quality control, providing an important basis for the subsequent formulation of multi-layer, multi-pass automatic routing strategies. The acquired three-dimensional coordinates of the weld bevel feature points and their corresponding welding torch posture parameters are transmitted in real time to the welding robot end effector for verification. The welding wire tip can accurately point to the identified feature points, and the absolute error between its three-dimensional coordinates and the theoretical value does not exceed 0.5 mm, meeting the requirements of high-precision welding operations.

[0032] Preferably, the formula for calculating the width of the root pass weld surface is as follows:

[0033]

[0034] The formula for calculating the cross-sectional area of ​​the bevel is as follows:

[0035]

[0036] Among them, subscript Indicates that the current frame is the [number]. frame, The width of the root pass weld surface in the 3D point cloud. The cross-sectional area of ​​the bevel. These are the intersection points of the branch pipe face and the bevel sidewall, the intersection point of the bevel sidewall and the root pass weld surface, the intersection point of the root pass weld surface and the main pipe face, and the intersection point of the branch pipe face and the main pipe face, respectively. If intersection point A is subtracted from intersection point B, then a vector is formed pointing from intersection point B to intersection point A. It is the modulus of orientation quantity. It is the cross product of vectors.

[0037] Preferably, the formula for calculating the bevel angle is as follows:

[0038]

[0039] The formula for calculating the angle between the bevel sidewall and the root pass surface is as follows:

[0040]

[0041] The formula for calculating the angle between the main weld surface and the root pass weld surface is as follows:

[0042]

[0043] Among them, subscript Indicates that the current frame is the [number]. frame, For the bevel angle, For the bevel sidewall and the root pass weld surface, The angle between the main weld surface and the root pass weld surface. These are the intersection points of the branch pipe face and the bevel sidewall, the intersection point of the bevel sidewall and the root pass weld surface, the intersection point of the root pass weld surface and the main pipe face, the intersection point of the branch pipe face and the main pipe face, and the virtual intersection point of the extension line of the bevel sidewall and the main pipe face. If intersection point A is subtracted from intersection point B, then a vector is formed from intersection point B to intersection point A. It is the modulus of orientation quantity. For vector dot product, It is the cross product of vectors.

[0044] Preferably, the bevel depth calculation process is as follows:

[0045] Calculate the direction vector of the bevel angle bisector of the saddle-shaped weld. ;

[0046] Construct the parametric equations for the bevel angle bisectors of the saddle-shaped weld: ;

[0047] By combining the three-dimensional point clouds of the bevel angle bisector and the root pass weld surface, the intersection point of the angle bisector and the fitted line of the root pass weld surface can be obtained. ;

[0048] By combining the three-dimensional point clouds of the bevel angle bisector and the branch pipe surface, the intersection point of the angle bisector and the fitted line of the branch pipe surface can be obtained. ;

[0049] Calculate bevel depth ;

[0050] Among them, subscript Indicates that the current frame is the [number]. frame, These are the intersection points of the bevel sidewall and the root pass surface, the intersection point of the root pass surface and the main pipe surface, and the virtual intersection point of the extension line of the bevel sidewall and the main pipe surface, respectively. If intersection point A is subtracted from intersection point B, then a vector is formed pointing from intersection point B to intersection point A. It is the modulus of orientation quantity. For parameters.

[0051] Preferably, the method further includes:

[0052] (1) Calculate the unit vector of the welding torch direction. :

[0053]

[0054]

[0055] (2) Calculate the welding direction vector :

[0056]

[0057] (3) The unit vector in the direction of the welding torch Z-axis vector of the robot tool coordinate system , with welding direction vector As the X-axis vector of the robot tool coordinate system, the cross product of the two vectors yields the Y-axis of the robot tool coordinate system. ;

[0058] (4) Construct the orthogonal rotation matrix of the robot tool coordinate system :

[0059]

[0060] (5) By using an orthogonal rotation matrix First calculate the rotation angle around the Y-axis. Then calculate the rotation angle around the Z-axis. and rotation angle around the X-axis Obtain the robot's posture during welding. ;

[0061] Among them, subscript Indicates that the current frame is the [number]. frame, Let be the welding torch direction vector. The angle between the welding torch and the bevel sidewall. These are the intersections of the branch pipe face and the bevel sidewall, the intersection of the bevel sidewall and the root pass weld surface, and the intersection of the root pass weld surface and the main pipe face, respectively. If intersection A is subtracted from intersection B, then a vector is formed pointing from intersection B to intersection A. It is the modulus of orientation quantity. For vector dot product, It is the cross product of vectors.

[0062] To achieve the above objectives, in a second aspect, this application provides a system for extracting and measuring the bevel features and dimensions of a thick plate saddle-shaped weld, comprising: at least one memory for storing a program; and at least one processor for accessing the program stored in the memory, wherein when the program stored in the memory is accessed, the processor is used to access the method described in the first aspect.

[0063] It is understandable that the beneficial effects of the second aspect mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here.

[0064] Overall, the technical solutions conceived in this application have the following beneficial effects compared with the prior art:

[0065] This application proposes a method for feature extraction and size measurement of saddle-shaped weld grooves in thick plates, including: acquiring a three-dimensional point cloud of the saddle-shaped weld groove in a thick plate; based on multi-scale geometric feature mutation detection, segmenting the three-dimensional point cloud of the weld groove in the current frame into four regions: the branch pipe surface, the groove sidewall, the root pass weld surface, and the main pipe surface, obtaining four regions of three-dimensional point clouds; determining the intersection points of the branch pipe surface and the groove sidewall, the intersection points of the groove sidewall and the root pass weld surface, the intersection points of the root pass weld surface and the main pipe surface, the intersection points of the branch pipe surface and the main pipe surface, and the virtual intersection points of the extension lines of the groove sidewall and the main pipe surface, which together serve as the feature points of the weld groove; constructing vectors between feature points, and calculating the geometric dimensions of the weld groove through vector operations. This application utilizes adaptive point cloud segmentation based on multi-scale geometric feature mutation detection to divide the spatial curve fitting of saddle-shaped weld grooves into segmented intervals. It does not require a preset slope threshold for the fitting curve, thus overcoming the limitations of traditional measurement methods on specific angle conditions. It can adapt to different angle conditions between branch pipes and main pipes in pipe-to-pipe intersection nodes, and is suitable for feature point identification and extraction and groove geometric dimension measurement of variable cross-section saddle-shaped weld grooves with varying groove contours and dimensions in the intersection space. Attached Figure Description

[0066] Figure 1 This is a flowchart of a method for extracting features and measuring dimensions of a saddle-shaped weld groove in a thick plate, provided in an embodiment of this application.

[0067] Figure 2 This is a schematic diagram of the three-dimensional point cloud of the saddle-shaped weld bevel and the saddle-shaped weld groove provided in the embodiments of this application.

[0068] Figure 3 This is a schematic diagram of the three-dimensional point cloud segmentation of the saddle-shaped weld groove contour provided in the embodiments of this application.

[0069] Figure 4 This is a schematic diagram of the three-dimensional point cloud profile and fitting curve of the j-th frame of the laser-scanned spatial intersecting variable cross-section saddle-shaped weld groove.

[0070] Figure 5 This is a schematic diagram of the three-dimensional point cloud feature points of the j-th frame of the laser-scanned spatial intersecting variable cross-section saddle-shaped weld groove provided in the embodiments of this application.

[0071] Figure 6 This is a schematic diagram of the included angles at various positions of the laser scanning spatial intersecting variable cross-section saddle-shaped weld groove in the j-th frame provided in the embodiments of this application.

[0072] Figure 7 This is a schematic diagram of the verification of the feature points of the saddle-shaped weld groove and the welding torch posture provided in the embodiments of this application.

[0073] Figure 8 This is a schematic diagram of the multi-layer, multi-pass welding result of the saddle-shaped weld groove provided in the embodiments of this application. Detailed Implementation

[0074] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0075] In this application, the term "and / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent three cases: A existing alone, A and B existing simultaneously, and B existing alone. In this application, the symbol " / " indicates that the related objects are in an "or" relationship, for example, A / B means A or B.

[0076] In this application, the terms "first" and "second," etc., are used to distinguish different objects, not to describe a specific order of objects. For example, "first response message" and "second response message," etc., are used to distinguish different response messages, not to describe a specific order of response messages.

[0077] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is 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 designs. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0078] In the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more, for example, multiple processing units means two or more processing units, multiple elements means two or more elements, etc.

[0079] The embodiments of this application are described below with reference to the accompanying drawings.

[0080] like Figure 1 As shown, this application provides a method for extracting features and measuring dimensions of a saddle-shaped weld groove in a thick plate, including:

[0081] Obtain the 3D point cloud of the saddle-shaped weld groove in a thick plate;

[0082] Based on multi-scale geometric feature mutation detection, the current frame weld groove three-dimensional point cloud is segmented into branch pipe surface, groove sidewall, root pass surface and main pipe surface, resulting in four regions of three-dimensional point cloud;

[0083] Based on the three-dimensional point cloud of each region, the intersection points of the branch pipe face and the bevel sidewall, the intersection points of the bevel sidewall and the root pass weld surface, the intersection points of the root pass weld surface and the main pipe face, the intersection points of the branch pipe face and the main pipe face, and the virtual intersection points of the bevel sidewall and the extension line of the main pipe face are determined and used together as the weld bevel feature points.

[0084] Construct vectors between feature points, and calculate the geometric dimensions of the weld groove through vector operations.

[0085] Preferably, the multi-scale geometric feature abrupt change detection divides the current frame weld groove 3D point cloud into four regions: branch pipe surface, groove sidewall, root pass weld surface, and main pipe surface, as detailed below:

[0086] Principal component analysis is performed on the neighborhood point set of each point in the current frame's 3D point cloud to obtain the eigenvalues ​​and corresponding eigenvectors of the covariance matrix in the principal component analysis.

[0087] The eigenvector corresponding to the smallest eigenvalue is taken as a point. Normal vector at the weld bevel ;

[0088] Calculate the proportion of the smallest eigenvalue to the sum of all eigenvalues, and use this proportion as the point. Curvature at the weld bevel ;

[0089] Calculate the change in the angle between the normal vectors of adjacent points ,in, For vector dot product, To take the absolute value;

[0090] Calculate curvature difference ;

[0091] Constructing a comprehensive feature function , These are the weighting coefficients, ;

[0092] like Judgment point This is the local maximum value point at the weld bevel.

[0093] Sort all local maximum points in descending order of comprehensive characteristic value, and retain the three largest ones as the dividing points;

[0094] The three detected segmentation points are indexed as segmentation point locations, and the three-dimensional point cloud of the weld groove is divided into four three-dimensional point clouds.

[0095] Preferably, the process of determining the characteristic points of the weld groove is as follows:

[0096] (1) The intersection of the branch pipe face and the bevel sidewall, the intersection of the bevel sidewall and the root pass weld surface, and the intersection of the root pass weld surface and the main pipe face are respectively represented as the length of the fitting window correction.

[0097] (2) Based on the intersection of the three-dimensional point cloud of each region and the above-mentioned intersection with the fitted window correction length, determine the intersection of the pipe surface and the bevel sidewall, the intersection of the bevel sidewall and the root pass weld surface, the intersection of the root pass weld surface and the main pipe surface, the intersection of the branch pipe surface and the main pipe surface, and the virtual intersection of the bevel sidewall and the extension line of the main pipe surface.

[0098] (3) Calculate the minimum distance deviation between the current frame's 3D point cloud and the four intersection points determined in this round;

[0099] (4) If the minimum distance deviation is greater than the preset value, adjust at least one fitting window correction length value and return to step (2); otherwise, take all current intersection points as feature points.

[0100] Preferably, adjusting at least one fitting window correction length value specifically involves:

[0101] If the minimum distance deviation corresponds to the bevel sidewall, then fine-tune the fitting window correction length of the intersection of the bevel sidewall and the fitting window correction length of the intersection of the bevel sidewall and the root pass surface to make the bevel surface point cloud more coplanar.

[0102] If the distance between the intersection of the bevel sidewall and the root pass surface and the intersection of the root pass surface and the main pipe surface is greater than a preset threshold, then the fitting window correction length of the intersection of the bevel sidewall and the root pass surface and the fitting window correction length of the intersection of the root pass surface and the main pipe surface are finely adjusted so that the distance between the intersection of the bevel sidewall and the root pass surface and the intersection of the root pass surface and the main pipe surface does not exceed the theoretical value range of the saddle-shaped weld gap weld width.

[0103] Preferably, the weld groove geometry includes: the width of the root pass weld surface, the groove depth, the groove angle, the groove width, the groove cross-sectional area, the angle between the groove sidewall and the root pass weld surface, and the angle between the main pipe surface and the root pass weld surface.

[0104] Preferably, the formula for calculating the width of the root pass weld surface is as follows:

[0105]

[0106] The formula for calculating the cross-sectional area of ​​the bevel is as follows:

[0107]

[0108] Among them, subscript Indicates that the current frame is the [number]. frame, The width of the root pass weld surface in the 3D point cloud. The cross-sectional area of ​​the bevel. These are the intersection points of the branch pipe face and the bevel sidewall, the intersection point of the bevel sidewall and the root pass weld surface, the intersection point of the root pass weld surface and the main pipe face, and the intersection point of the branch pipe face and the main pipe face, respectively. If intersection point A is subtracted from intersection point B, then a vector is formed pointing from intersection point B to intersection point A. It is the modulus of orientation quantity. It is the cross product of vectors.

[0109] Preferably, the formula for calculating the bevel angle is as follows:

[0110]

[0111] The formula for calculating the angle between the bevel sidewall and the root pass surface is as follows:

[0112]

[0113] The formula for calculating the angle between the main weld surface and the root pass weld surface is as follows:

[0114]

[0115] Among them, subscript Indicates that the current frame is the [number]. frame, For the bevel angle, For the bevel sidewall and the root pass weld surface, The angle between the main weld surface and the root pass weld surface. These are the intersection points of the branch pipe face and the bevel sidewall, the intersection point of the bevel sidewall and the root pass weld surface, the intersection point of the root pass weld surface and the main pipe face, the intersection point of the branch pipe face and the main pipe face, and the virtual intersection point of the extension line of the bevel sidewall and the main pipe face. If intersection point A is subtracted from intersection point B, then a vector is formed from intersection point B to intersection point A. It is the modulus of orientation quantity. For vector dot product, It is the cross product of vectors.

[0116] Preferably, the bevel depth calculation process is as follows:

[0117] Calculate the direction vector of the bevel angle bisector of the saddle-shaped weld. ;

[0118] Construct the parametric equations for the bevel angle bisectors of the saddle-shaped weld: ;

[0119] By combining the three-dimensional point clouds of the bevel angle bisector and the root pass weld surface, the intersection point of the angle bisector and the fitted line of the root pass weld surface can be obtained. ;

[0120] By combining the three-dimensional point clouds of the bevel angle bisector and the branch pipe surface, the intersection point of the angle bisector and the fitted line of the branch pipe surface can be obtained. ;

[0121] Calculate bevel depth ;

[0122] Among them, subscript Indicates that the current frame is the [number]. frame, These are the intersection points of the bevel sidewall and the root pass surface, the intersection point of the root pass surface and the main pipe surface, and the virtual intersection point of the extension line of the bevel sidewall and the main pipe surface, respectively. If intersection point A is subtracted from intersection point B, then a vector is formed pointing from intersection point B to intersection point A. It is the modulus of orientation quantity. For parameters.

[0123] Preferably, the method further includes:

[0124] (1) Calculate the unit vector of the welding torch direction. :

[0125]

[0126]

[0127] (2) Calculate the welding direction vector :

[0128]

[0129] (3) The unit vector in the direction of the welding torch Z-axis vector of the robot tool coordinate system , with welding direction vector As the X-axis vector of the robot tool coordinate system, the cross product of the two vectors yields the Y-axis of the robot tool coordinate system. ;

[0130] (4) Construct the orthogonal rotation matrix of the robot tool coordinate system :

[0131]

[0132] (5) By using an orthogonal rotation matrix First calculate the rotation angle around the Y-axis. Then calculate the rotation angle around the Z-axis. and rotation angle around the X-axis Obtain the robot's posture during welding. ;

[0133] Among them, subscript Indicates that the current frame is the [number]. frame, Let be the welding torch direction vector. The angle between the welding torch and the bevel sidewall. These are the intersections of the branch pipe face and the bevel sidewall, the intersection of the bevel sidewall and the root pass weld surface, and the intersection of the root pass weld surface and the main pipe face, respectively. If intersection A is subtracted from intersection B, then a vector is formed pointing from intersection B to intersection A. It is the modulus of orientation quantity. For vector dot product, It is the cross product of vectors.

[0134] Example

[0135] 1. Laser vision scanning to acquire 3D point cloud of spatially intersecting, multi-layered, multi-pass variable cross-section saddle-shaped weld groove. , To determine the number of points in the 3D point cloud when scanning a frame of a saddle-shaped weld groove profile, To scan the maximum number of frames in a point cloud dataset, such as Figure 2 As shown.

[0136] 2. With the first The frame contains 3D point cloud data for the current frame, and instructions for subsequent operations are provided.

[0137]

[0138] in, Indices representing the indexes of three-dimensional points. , The index representing the scan frame. .

[0139] 3. Divide the point cloud into four regions: branch pipe surface, bevel sidewall, root pass weld surface, and main pipe surface. Figure 3 As shown.

[0140] For the Frame 3D point cloud Each point in Principal Component Analysis (PCA) is performed on the neighborhood point set of the normal vector. The eigenvector corresponding to the smallest eigenvalue is the normal vector. The eigenvalues ​​of the neighborhood point set covariance matrix are sorted in ascending order: Calculate the proportion of the smallest eigenvalue to the sum of all eigenvalues, and use this proportion as the point. The curvature is It reflects the "flatness" of the point cloud along the direction of the minimum eigenvalue in the local neighborhood: when this value is large, it indicates that the local neighborhood changes drastically in a certain direction (i.e., large curvature); when this value is small, it indicates that the local neighborhood is close to a plane (i.e., small curvature).

[0141] Generate feature sequence: Calculate the change in the angle between the normal vectors of adjacent points: Calculate curvature difference .

[0142] Mutation point detection: Constructing a comprehensive feature function: , These are the weighting coefficients. .

[0143] Peak detection: smoothing Detecting local maximum points ,and , This represents the empirical value for peak detection.

[0144] Screening for key mutation points: Retain the 3 peak points with the highest significance (according to...) (Take the first 3 values ​​in descending order).

[0145] Minimum Spacing Constraint: Distance between adjacent split points ,in, This is the theoretical minimum width of the base surface.

[0146] Region segmentation: The three detected segmentation points are indexed as the segmentation point locations. The point cloud was divided into four regions: the branch pipe surface, the bevel sidewall, the root pass weld surface, and the main pipe surface. ,like Figure 4 As shown.

[0147] 4. Define the characteristic points of the saddle-shaped weld groove as... ,like Figure 5 As shown, the coordinates of each feature point are: , For the first The point cloud coordinates of each point.

[0148] 5. Set the point cloud for each region.

[0149] branch section ;

[0150] Bevel sidewall section ;

[0151] base section ;

[0152] Supervisor Section .

[0153] 6. Fit geometric models for each of the four region point clouds.

[0154] Branch pipe sidewall → cylindrical surface, bevel sidewall → flat surface, root pass weld flat surface → flat surface, main pipe sidewall → cylindrical surface.

[0155] 7. Settings The empirical value for adjusting the fitting window length for each feature point is related to the number of curve points where the feature point is located.

[0156] 8. Intersection of branch pipe face and bevel sidewall .

[0157] 9. Intersection of the bevel sidewall and the base surface .

[0158] 10. Intersection of the base surface and the side wall of the main pipe .

[0159] 11. Intersection of the branch pipe sidewall and the main pipe sidewall .

[0160] 12. Intersection of the extended lines of the bevel sidewall and the main pipe sidewall .

[0161] The first four are all intersections of entities. This is a virtual intersection.

[0162] 13. Calculate 3D point clouds Intersection with 4 points Minimum distance deviation between :

[0163]

[0164] 14. Residual-based adaptive window adjustment, if This indicates a bias in the fitted feature points, so the fitting window length should be adjusted. Value, repeat steps 8-13; if Then feature points are preserved. , These are empirical values ​​and are related to the tolerance accuracy.

[0165] Specifically, the minimum error Adjust the position of the dividing point: If Large errors in the bevel sidewalls require fine-tuning. and Make the bevel face more uniform; if and Too wide a distance Fine-tuning and This makes the minimum length of the root pass weld surface < "Maximum length of the root pass weld surface".

[0166] 15. Calculate the geometric dimensions of the bevel, such as... Figure 6 As shown.

[0167] bevel angle ;

[0168] Angle between the bevel sidewall and the root pass surface ;

[0169] Angle between the side wall of the main pipe and the surface of the root pass weld .

[0170] 16. Direction vector of the bevel angle bisector of the saddle-shaped weld. .

[0171] Parameter equations for the angle bisector of the saddle-shaped weld groove , For parameters.

[0172] Joint slope corner dividing line 3D point cloud of the root pass weld surface The intersection point of the angle bisector and the fitted line of the root pass weld surface is obtained by solving. .

[0173] Joint slope corner dividing line 3D point cloud of branch pipe face The intersection point of the angle bisector and the fitted line of the branch pipe surface is obtained by solving. .

[0174] 17. Calculate the root bevel width of the saddle-shaped weld. .

[0175] 18. Calculate the groove depth of the saddle-shaped weld. .

[0176] 19. Cross-sectional area of ​​saddle-shaped weld groove .

[0177] 20. Set the welding start point position as Set the angle between the welding torch and the bevel sidewall as follows: , The size is set by the welding process requirements and is an empirical value.

[0178] Calculate the welding torch direction vector ;

[0179] .

[0180] 21. Calculate the welding direction of the saddle-shaped weld seam as the robot tool coordinate system. , with welding torch direction vector Robot tool coordinate system , and Cross product yields the tool coordinate system. .

[0181]

[0182]

[0183]

[0184] Constructing the orthogonal rotation matrix of the robot tool coordinate system :

[0185]

[0186] Solve the rotation angle about the zyx axis robot welding posture .

[0187] First calculate the rotation angle around the Y-axis Then calculate the rotation angle around the Z-axis. and rotation angle around the X-axis The robot's posture during welding was obtained. .

[0188] 22. Constructing a welding robot in the first... Frame of saddle-shaped weld groove automated welding position and attitude information vector .

[0189] To verify the practical application effect of the proposed weld groove feature recognition method, this application transmits the three-dimensional coordinates of the obtained weld groove feature points and their corresponding welding torch pose parameters to the welding robot end effector in real time for verification (e.g., Figure 7 (As shown). Welding tests show that the welding wire tip can accurately point to the characteristic point at the junction of the root pass and the bevel sidewall, with the absolute error between its three-dimensional coordinates and the theoretical value not exceeding 0.5 mm, meeting the requirements of high-precision welding operations. Furthermore, the welding torch posture meets the requirements of subsequent welding processes, maintaining a preset working angle between its axis and the weld tangent direction, verifying the accuracy of the algorithm's recognition, and demonstrating that laser vision intelligent recognition technology based on point cloud dynamic segmentation can establish a precise geometric benchmark for automated welding. Weld formation is as follows... Figure 8 As shown, the weld formation quality is good and the surface is continuous and dense, indicating that the identification and measurement system has stable and reliable performance in the actual welding process and can meet the process requirements of multi-layer and multi-pass welding.

[0190] Comprehensive experimental results show that the method proposed in this application can effectively adapt to the geometric changes of saddle-shaped weld grooves caused by processing and assembly deviations in engineering, and realize feature recognition and accurate measurement of groove dimensions before welding, providing reliable data support and technical guidance for subsequent multi-layer and multi-pass automated welding.

[0191] It should be understood that the above-described device is used to execute the methods in the above embodiments. The implementation principle and technical effect of the corresponding program modules in the device are similar to those described in the above methods. The working process of the device can be referred to the corresponding process in the above methods, and will not be repeated here.

[0192] Based on the methods in the above embodiments, this application provides an electronic device that may include a processor, a communications interface, a memory, and a communication bus, wherein the processor, communications interface, and memory communicate with each other via the communication bus. The processor may invoke logical instructions stored in the memory to execute the methods in the above embodiments.

[0193] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0194] Based on the methods in the above embodiments, this application provides a computer-readable storage medium storing a computer program that, when run on a processor, causes the processor to execute the methods in the above embodiments.

[0195] Based on the methods in the above embodiments, this application provides a computer program product that, when run on a processor, causes the processor to execute the methods in the above embodiments.

[0196] It is understood that the processor in the embodiments of this application can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. A general-purpose processor can be a microprocessor or any conventional processor.

[0197] The method steps in this application embodiment can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in random access memory (RAM), flash memory, read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), registers, hard disks, portable hard disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can reside in an ASIC.

[0198] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted through the computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0199] It is understood that the various numerical designations used in the embodiments of this application are merely for the convenience of description and are not intended to limit the scope of the embodiments of this application.

[0200] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for extracting features and measuring dimensions of a saddle-shaped weld groove in a thick plate, characterized in that, include: Obtain the 3D point cloud of the saddle-shaped weld groove in a thick plate; Based on multi-scale geometric feature mutation detection, the current frame weld groove three-dimensional point cloud is segmented into branch pipe surface, groove sidewall, root pass surface and main pipe surface, resulting in four regions of three-dimensional point cloud; Based on the three-dimensional point cloud of each region, the intersection points of the branch pipe face and the bevel sidewall, the intersection points of the bevel sidewall and the root pass weld surface, the intersection points of the root pass weld surface and the main pipe face, the intersection points of the branch pipe face and the main pipe face, and the virtual intersection points of the bevel sidewall and the extension line of the main pipe face are determined and used together as the weld bevel feature points. Construct vectors between feature points, and calculate the geometric dimensions of the weld groove through vector operations; Specifically, the multi-scale geometric feature mutation detection method segments the current frame weld groove 3D point cloud into four regions: branch pipe surface, groove sidewall, root pass weld surface, and main pipe surface, as detailed below: Principal component analysis is performed on the neighborhood point set of each point in the current frame's 3D point cloud to obtain the eigenvalues ​​and corresponding eigenvectors of the covariance matrix in the principal component analysis. The eigenvector corresponding to the smallest eigenvalue is taken as a point. Normal vector at the weld bevel ; Calculate the proportion of the smallest eigenvalue to the sum of all eigenvalues, and use this proportion as the point. Curvature at the weld bevel ; Calculate the change in the angle between the normal vectors of adjacent points ,in, For vector dot product, To take the absolute value; Calculate curvature difference ; Constructing a comprehensive feature function , These are the weighting coefficients, ; like Judgment point This is the local maximum value point at the weld bevel. Sort all local maximum points in descending order of comprehensive characteristic value, and retain the three largest ones as the dividing points; The three detected segmentation points are indexed as segmentation point locations, and the three-dimensional point cloud of the weld groove is divided into four three-dimensional point clouds. The process of determining the characteristic points of the weld groove is as follows: (1) The intersection of the branch pipe face and the bevel sidewall, the intersection of the bevel sidewall and the root pass weld surface, and the intersection of the root pass weld surface and the main pipe face are respectively represented as the length of the fitting window correction. (2) Based on the intersection of the three-dimensional point cloud of each region and the above-mentioned intersection with the fitted window correction length, determine the intersection of the pipe surface and the bevel sidewall, the intersection of the bevel sidewall and the root pass weld surface, the intersection of the root pass weld surface and the main pipe surface, the intersection of the branch pipe surface and the main pipe surface, and the virtual intersection of the bevel sidewall and the extension line of the main pipe surface. (3) Calculate the minimum distance deviation between the current frame's 3D point cloud and the four intersection points determined in this round; (4) If the minimum distance deviation is greater than the preset value, adjust at least one fitting window correction length value and return to step (2); otherwise, take all current intersection points as feature points.

2. The method as described in claim 1, characterized in that, The adjustment of at least one fitting window correction length value specifically includes: If the minimum distance deviation corresponds to the bevel sidewall, then fine-tune the fitting window correction length of the intersection of the bevel sidewall and the fitting window correction length of the intersection of the bevel sidewall and the root pass surface to make the bevel surface point cloud more coplanar. If the distance between the intersection of the bevel sidewall and the root pass surface and the intersection of the root pass surface and the main pipe surface is greater than a preset threshold, then the fitting window correction length of the intersection of the bevel sidewall and the root pass surface and the fitting window correction length of the intersection of the root pass surface and the main pipe surface are finely adjusted so that the distance between the intersection of the bevel sidewall and the root pass surface and the intersection of the root pass surface and the main pipe surface does not exceed the theoretical value range of the saddle-shaped weld gap weld width.

3. The method as described in claim 1, characterized in that, The geometric dimensions of the weld bevel include: the width of the root pass weld surface, the bevel depth, the bevel angle, the bevel width, the bevel cross-sectional area, the angle between the bevel sidewall and the root pass weld surface, and the angle between the main pipe surface and the root pass weld surface.

4. The method as described in claim 3, characterized in that, The formula for calculating the width of the root pass weld surface is as follows: The formula for calculating the cross-sectional area of ​​the bevel is as follows: Among them, subscript Indicates that the current frame is the [number]. frame, The width of the root pass weld surface in the 3D point cloud. The cross-sectional area of ​​the bevel. These are the intersection points of the branch pipe face and the bevel sidewall, the intersection point of the bevel sidewall and the root pass weld surface, the intersection point of the root pass weld surface and the main pipe face, and the intersection point of the branch pipe face and the main pipe face, respectively. If intersection point A is subtracted from intersection point B, then a vector is formed pointing from intersection point B to intersection point A. It is the magnitude of the orientation quantity. It is the cross product of vectors.

5. The method as described in claim 3, characterized in that, The formula for calculating the bevel angle is as follows: The formula for calculating the angle between the bevel sidewall and the root pass surface is as follows: The formula for calculating the angle between the main weld surface and the root pass weld surface is as follows: Among them, subscript Indicates that the current frame is the [number]. frame, For the bevel angle, The angle between the bevel sidewall and the root pass surface. The angle between the main weld surface and the root pass weld surface. These are the intersection points of the branch pipe face and the bevel sidewall, the intersection point of the bevel sidewall and the root pass weld surface, the intersection point of the root pass weld surface and the main pipe face, the intersection point of the branch pipe face and the main pipe face, and the virtual intersection point of the extension line of the bevel sidewall and the main pipe face. If intersection point A is subtracted from intersection point B, then a vector is formed from intersection point B to intersection point A. It is the magnitude of the orientation quantity. It is the dot product of vectors.

6. The method as described in claim 3, characterized in that, The process for calculating the bevel depth is as follows: Calculate the direction vector of the bevel angle bisector of the saddle-shaped weld. ; Construct the parametric equations for the bevel angle bisectors of the saddle-shaped weld: ; By combining the three-dimensional point clouds of the bevel angle bisector and the root pass weld surface, the intersection point of the angle bisector and the fitted line of the root pass weld surface can be obtained. ; By combining the three-dimensional point clouds of the bevel angle bisector and the branch pipe surface, the intersection point of the angle bisector and the fitted line of the branch pipe surface can be obtained. ; Calculate bevel depth ; Among them, subscript Indicates that the current frame is the [number]. frame, These are the intersection points of the bevel sidewall and the root pass surface, the intersection point of the root pass surface and the main pipe surface, and the virtual intersection point of the extension line of the bevel sidewall and the main pipe surface, respectively. If intersection point A is subtracted from intersection point B, then a vector is formed pointing from intersection point B to intersection point A. It is the magnitude of the orientation quantity. For parameters.

7. The method according to any one of claims 1 to 6, characterized in that, The method also includes: (1) Calculate the unit vector of the welding torch direction. : (2) Calculate the welding direction vector : (3) The unit vector in the direction of the welding torch Z-axis vector of the robot tool coordinate system , with welding direction vector As the X-axis vector of the robot tool coordinate system, the cross product of the two vectors yields the Y-axis of the robot tool coordinate system. ; (4) Construct the orthogonal rotation matrix of the robot tool coordinate system : (5) By using an orthogonal rotation matrix First calculate the rotation angle around the Y-axis. Then calculate the rotation angle around the Z-axis. and rotation angle around the X-axis Obtain the robot's posture during welding. ; Among them, subscript Indicates that the current frame is the [number]. frame, Let be the welding torch direction vector. The angle between the welding torch and the bevel sidewall. These are the intersections of the branch pipe face and the bevel sidewall, the intersection of the bevel sidewall and the root pass weld surface, and the intersection of the root pass weld surface and the main pipe face, respectively. If intersection A is subtracted from intersection B, then a vector is formed pointing from intersection B to intersection A. It is the magnitude of the orientation quantity. It is the cross product of vectors.

8. A system for extracting features and measuring dimensions of saddle-shaped weld grooves in thick plates, characterized in that, include: At least one memory for storing programs; At least one processor is configured to access a program stored in the memory, wherein, when the program stored in the memory is accessed, the processor is configured to access the method as described in any one of claims 1 to 7.

Images