Method for improving positioning compatibility of automatic welding seam position

Abstract

The invention discloses a method for improving the positioning compatibility of an automatic welding seam position, which comprises the following steps of: performing hand-eye calibration by using a high-precision 3D camera and a welding robot, and determining a pose matrix between a camera coordinate system and a robot geodetic coordinate system; obtaining a pose matrix of an actual workpiece in a robot geodetic coordinate system through methods such as point cloud matching; a welding seam local coordinate system is established through welding seam mathematical analysis; acquiring actual workpiece point cloud data; generating a registration template of the point cloud model of the linear welding seam by using the steel member digital model; the positions of a weld joint starting point and a weld joint ending point in the actual point cloud under a camera coordinate system are obtained; and converting coordinates of a welding seam starting point and a welding seam ending point under a camera coordinate system into a robot geodetic coordinate system, and welding a workpiece by matching with an expert welding process library. The method has wide applicability and high universality, steel member welding seams of various models and sizes can be rapidly and accurately positioned, and then the efficiency and reliability of the automatic welding process are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of industrial intelligent manufacturing, and in particular relates to a method for improving the compatibility of automated welding weld position positioning. Background Art

[0002] In modern industrial production, automated welding technology is a key technology to improve production efficiency and ensure welding quality. However, due to the diversity and complexity of steel components, traditional automated welding technology has limitations in weld positioning. In addition, traditional automated welding technology usually uses 3D vision to guide welding robots for welding, and visual algorithms need to be frequently adjusted and optimized when facing steel components of different models and specifications. The compatibility is not high, resulting in low work efficiency. Summary of the invention

[0003] In order to solve the technical problems existing in the prior art, the object of the present invention is to provide a method for improving the compatibility of automated welding weld position positioning.

[0004] In order to achieve the above purpose and the above technical effect, the technical solution adopted by the present invention is:

[0005] A method for improving the compatibility of automated welding seam positioning comprises the following steps:

[0006] Step 1: Use a high-precision 3D camera and welding robot to perform hand-eye calibration and establish the pose matrix M1 between the camera coordinate system and the robot's earth coordinate system;

[0007] Step 2: Obtain the pose matrix M2 of the actual workpiece in the robot's earth coordinate system;

[0008] Step 3: Establish the local coordinate system of the weld through the weld digital model analysis;

[0009] Step 4: Obtain actual workpiece point cloud data;

[0010] Step 5: Generate a registration template for the point cloud model of the linear weld using the steel component digital model;

[0011] Step 6: Get the positions of the starting point and end point of the weld in the actual point cloud in the camera coordinate system;

[0012] Step 7: Transform the coordinates of the weld start and end points in the camera coordinate system in step 6 to the robot earth coordinate system, and weld the workpiece using the expert welding process library.

[0013] Furthermore, in step 2, the step of obtaining the pose matrix M2 of the actual workpiece in the robot's earth coordinate system includes:

[0014] 21) Touch at least three points on the actual workpiece with the tip of the welding gun tool. The distance between the points should be as evenly distributed as possible on the workpiece to improve the calculation accuracy, and obtain the coordinate value (X, Y, Z) of the current robot tool TCP point;

[0015] 22) Read the model point coordinate (X, Y, Z) value corresponding to the robot TCP point in step 21) in the 3D model software;

[0016] 23) Solve the pose matrix M2 through the two sets of corresponding points obtained in step 21) and step 22).

[0017] Furthermore, in step 2, the step of obtaining the pose matrix M2 of the actual workpiece in the robot's earth coordinate system includes:

[0018] The entire workpiece is photographed using a large-field line scan camera, and the pose matrix M2 is calculated using the 3D registration algorithm in combination with the workpiece digital model.

[0019] Furthermore, in step 3, the steps of establishing the weld local coordinate system by analyzing the weld digital model include:

[0020] 31) Analyze the digital model of the steel structure to be welded and obtain the standard position information of the weld;

[0021] 32) Establish the starting and ending coordinates of the weld;

[0022] 33) The weld position and the normal vectors of the two side surfaces are transformed to the robot earth coordinate system through the pose matrix M2, and the weld coordinate system is established using the transformed weld position normal vector to represent the position of the weld in the robot earth coordinate system.

[0023] Furthermore, the weld includes a flat weld and a vertical weld, and the judgment principle of the flat weld and the vertical weld is:

[0024] According to the angle between the weld normal vector and the Z-axis normal vector of the robot's geodetic coordinate system, an angle >70° is a flat weld, and an angle <30° is a vertical weld.

[0025] Furthermore, the steps for establishing the coordinate system at the starting point of the flat weld include:

[0026] 311) The normal vectors of the two sides of the weld F1 and F2 are known. The two end points of the weld, A and B;

[0027] Calculate the angle between the normal vectors of faces F1 and F2 and the Z-axis normal vector of the world coordinate system to determine the Z-axis, take the side with the smallest angle, and use the surface normal vector as the basis vector of the Z-axis

[0028] 312) Use Cross product weld vector or Right now or is the normal vector of the faces standing on both sides of the weld;

[0029] 313) Use Multiply the normal vector of the other face by the dot to determine the angle. If the angle is less than 10°, then the vector or As the basis vector of the Y axis of the coordinate system;

[0030] 314) According to the vector Determine the X-axis basis vector;

[0031] 315) Determine the vector that meets the condition of step 312) or Use point A or point B as the origin of the coordinate system; this calculation method is not effective when the angle between surface F1 and surface F2 is not 90°. The vertical plane of the two faces that are not parallel to the weld;

[0032] 316) Establish the weld end point coordinate system: Replace the origin of the coordinate system calculated in the first five steps with another point as the weld end point coordinate system.

[0033] Furthermore, the steps of establishing the coordinate system at the starting point of the vertical weld include:

[0034] 411) The normal vectors of the two sides of the weld F1 and F2 are known. The two end points of the weld, A and B;

[0035] The vector of the straight line connecting the start and end points of the weld or As the Z-axis basis vector of the coordinate system, determine or Is the normal vector of the world coordinate system in the same direction? If so, it is considered or As the Z-axis basis vector of the weld coordinate system At the same time, point A or point B is identified as the starting point of the weld, otherwise, point B or point A is identified as the end point of the weld;

[0036] 412) Use the normal vector of face F1 and the normal vector of face F2 Perform a cross product if The vector and If the axes are in the same direction, then As the X-axis basis vector, otherwise As the X-axis basis vector;

[0037] 413) According to the vector Determine the Y-axis basis vector;

[0038] 414) The weld start and end coordinates A or B determined in step 411) are used as the origin of the weld start or end coordinate system, and the weld coordinate system is established. If the angle between the surface F1 and the surface F2 is not 90°, then Not parallel to surface F2.

[0039] Furthermore, in step 4, the step of obtaining actual workpiece point cloud data includes:

[0040] Through the RRT* trajectory planning algorithm, the welding robot is moved to the vicinity of the actual workpiece;

[0041] The shooting posture Rx, Ry, and Rz of the high-precision 3D camera are set with an empirical value relative to the weld coordinate system. According to the camera field of view and anti-collision algorithm, the appropriate shooting distance d is calculated. The posture matrix M3 of the camera relative to the weld coordinate system is calculated based on Rx, Ry, Rz, and d.

[0042] Through the M1, M2, and M3 posture matrix transformations, the welding robot is moved to the weld coordinate system, and a high-precision 3D camera is used to take photos to obtain the actual workpiece point cloud data.

[0043] Furthermore, in step 5, the step of using the steel component digital model to generate a registration template of the point cloud model of the linear weld includes:

[0044] 51) The weld coordinate system is obtained from step 2, and the camera's field of view is known through factory parameters;

[0045] 52) The workpiece digital model is transformed into the camera coordinate system through the pose matrices M2 and M3;

[0046] 53) Define a virtual telecentric lens camera internal parameter;

[0047] 54) According to the camera internal parameters, the digital-analog mesh network coordinates within the field of view of the virtual telecentric lens camera are projected into the image coordinate system. At the same time, the three-dimensional coordinates of the corresponding points in the mesh network are recorded according to each pixel point of the image. The point cloud model within the field of view of the virtual telecentric lens camera is generated by the three-dimensional coordinates. If the corresponding points in multiple mesh networks are projected into the same image coordinates, the points closest to the camera coordinate system are retained; the size of the generated local model point cloud can be adjusted by adjusting the pixel size and the image width and height; the local model of the weld is cropped by using the predicted weld length information, and only the important parts of the weld and the surfaces on both sides of the weld are retained to reduce the interference of subsequent point cloud alignment.

[0048] Furthermore, in step 6, the step of obtaining the positions of the start point and the end point of the weld in the actual point cloud in the camera coordinate system includes:

[0049] The point cloud collected by the camera in step 4 is filtered, and based on the weld coordinate system, the actual point cloud is pre-transformed to the weld coordinate system. According to the weld length information, the actual point cloud is cropped to retain only important welds and surface information. Then, the actual point cloud is transformed back to the camera coordinate system through inverse transformation. Based on the high-precision point cloud registration algorithm, the homogeneous transformation matrix M4 from the point cloud model to the actual point cloud or the steel component point cloud collected by the camera is calculated, thereby transforming the starting and ending position coordinates of the weld in the steel construction model to obtain the positions of the starting and ending points of the weld in the actual point cloud in the camera coordinate system.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] The present invention discloses a method for improving the compatibility of weld position positioning in automated welding, comprising the following steps: using a high-precision 3D camera and a welding robot to perform hand-eye calibration to establish a pose matrix between a camera coordinate system and a robot earth coordinate system; obtaining a pose matrix of an actual workpiece in the robot earth coordinate system by point cloud matching and other methods; establishing a weld local coordinate system by weld digital model analysis; obtaining actual workpiece point cloud data; using a steel component digital model to generate a registration template for a point cloud model of a linear weld; obtaining the positions of a weld start point and an end point in the actual point cloud in the camera coordinate system; transforming the weld start point and end point coordinates in the camera coordinate system to the robot earth coordinate system, and welding the workpiece in combination with an expert welding process library; the present invention has wide applicability and high versatility, can quickly and accurately locate welds of steel components of various models and sizes, thereby improving the efficiency and reliability of the automated welding process, and solving the compatibility problem of visual algorithms in the prior art. DETAILED DESCRIPTION

[0052] The present invention is described in detail below so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby making a clearer and more definite definition of the protection scope of the present invention.

[0053] A brief summary of one or more aspects is given below to provide a basic understanding of these aspects. This summary is not an exhaustive overview of all conceived aspects, and is neither intended to identify the key or critical elements of all aspects nor to define the scope of any or all aspects. Its only purpose is to give some concepts of one or more aspects in a simplified form as a prelude to a more detailed description that will be given later.

[0054] A method for improving the compatibility of automated welding seam positioning comprises the following steps:

[0055] Step 1: Use a high-precision 3D camera and welding robot to perform hand-eye calibration and establish the pose matrix M1 between the camera coordinate system and the robot's earth coordinate system.

[0056] Step 2: Obtain the pose matrix M2 of the actual workpiece in the robot's earth coordinate system. There are two ways to obtain the pose matrix M2:

[0057] The first method is:

[0058] 21) Touch at least three points on the actual workpiece with the tip of the welding gun tool. The distance between the points should be as evenly distributed as possible on the workpiece to improve the calculation accuracy, and obtain the coordinate value (X, Y, Z) of the current robot tool TCP point;

[0059] 22) Read the model point coordinate (X, Y, Z) value corresponding to the robot TCP point in step 21) in the 3D model software;

[0060] 23) Solving the pose matrix M2 through the two sets of corresponding points obtained in step 21) and step 22);

[0061] The second method is to take a picture of the entire workpiece with a large field of view line scan camera, combine the workpiece digital model, and use the 3D registration algorithm to calculate the pose matrix M2.

[0062] Step 3: Establish the local coordinate system of the weld by analyzing the weld digital model. Usually in the actual production environment, only the digital models of steel components in STL, STP, Tekla format can be obtained. The digital models are generally provided by the manufacturer. By analyzing the digital model of the steel component to be welded, the standard weld position information is obtained, such as the three-dimensional coordinates of the starting point and end point of the weld, and the normal vectors of the two sides of the weld. The position coordinates can represent the position of the weld in the three-dimensional model of the steel component. The normal vector can provide direction information. The weld position and the normal vectors of the two sides are transformed to the robot's geodetic coordinate system through the M2 matrix. The weld coordinate system is established using the transformed weld position normal vector and other information. This coordinate system represents the position of the weld in the robot's geodetic coordinate system. The detailed steps are as follows:

[0063] First, the weld types are divided into flat welds and vertical welds. Flat welds and vertical welds are judged based on the angle between the weld normal vector and the normal vector of the Z axis of the world coordinate system. Angles > 70° are flat welds; angles < 30° are vertical welds.

[0064] Establishment of coordinate system at the starting point of flat weld: Normal vectors of two sides F1 and F2 of weld are known The two endpoints of the weld are A and B (it is unknown whether A or B is the starting point or the end point). Step 1: Calculate the angle between the normal vectors of faces F1 and F2 and the normal vector of the Z axis of the world coordinate system to determine the Z axis (method: calculate the angle through vector dot multiplication and take the face with the smallest angle). If the angle is less than a certain value, such as 10°, use the face normal vector as the basis vector of the Z axis. Step 2: Use Cross product weld vector or Right now (or ) is the normal vector of the faces standing on both sides of the weld. Multiply the normal vector of the other face by the dot to determine the angle. If the angle is less than 10°, then the vector (or ) as the basis vector of the Y axis of the coordinate system. Step 3: Based on the vector Determine the X-axis basis vector. Step 4: Determine the vector that meets the conditions in step 2 (or ), using point A or point B as the origin of the coordinate system. This calculation method is not effective when the angle between faces F1 and F2 is not 90°. The vertical plane of the two faces that are not parallel to the weld. Step 5: Establish the weld end point coordinate system: Replace the origin of the coordinate system calculated in the first four steps with another point (A or B) as the weld end point coordinate system.

[0065] Establishment of the coordinate system at the starting point of the vertical weld: the normal vectors of the two sides of the weld F1 and F2 are known The two endpoints of the weld are A and B (it is unknown whether A or B is the starting point or the end point). Step 1: The vector of the straight line connecting the starting point and the end point of the weld (or ) as the Z-axis basis vector of the coordinate system, determine (or ) and the normal vector of the world coordinate system are in the same direction. If they are consistent, then it is determined (or ) as the Z-axis basis vector of the weld coordinate system At the same time, identify point A (or point B) as the starting point of the weld, otherwise, identify point B (or point A) as the end point of the weld. Step 2: Use the normal vector of face F1 and the normal vector of face F2 Perform a cross product if The vector and If the axes are in the same direction, then As the X-axis basis vector, otherwise As the X-axis basis vector. Step 3: According to the vector Determine the Y-axis basis vector. Step 4: Use the weld start and end coordinates A or B determined in the first step as the origin of the weld start or end coordinate system, and the vertical weld coordinate system is now established. Note: If the angle between surface F1 and surface F2 is not 90°, then Not parallel to surface F2.

[0066] Step 4: Get the actual workpiece point cloud data. Through the RRT* trajectory planning algorithm, move the robot to the top of the actual workpiece; the camera's shooting posture Rx, Ry, Rz can be set with an empirical value relative to the weld coordinate system in step 2. According to the camera field of view and anti-collision algorithm, calculate the appropriate shooting distance d. According to Rx, Ry, Rz, d, the camera's posture matrix M3 relative to the weld coordinate system can be calculated; through the M1, M2, M3 matrix transformation, move the camera on the robot tool to the weld coordinate system, and the camera takes pictures to obtain the actual workpiece point cloud.

[0067] Step 5: Generate a registration template for the point cloud model of the linear weld using the steel component digital model. Since the size of the steel component is usually about 2 meters to 20 meters, in order to ensure accuracy, the point cloud sampling interval needs to be less than 0.5 mm, and the number of point clouds can reach the order of millions. The algorithm for generating a point cloud in ply format using the mesh network of the steel component digital model is very time-consuming and difficult to meet actual production needs. Therefore, the local point cloud registration template is generated by the following method: (1) The position of the weld coordinate system can be obtained from step 2, and the camera's field of view is known through factory parameters. (2) The workpiece digital model is transformed to the camera coordinate system through the M2 pose matrix in step 2 and the M3 pose matrix in step 3. (3) Define a virtual telecentric lens camera internal parameters such as: magnification M, distortion K, pixel size (Sx, Sy), image width and height (W, H), and image principal point (W / 2, H / 2). (4) According to the camera internal parameters, the digital model mesh network coordinates within the field of view of the virtual telecentric lens camera are projected into the image coordinate system. At the same time, the three-dimensional coordinates of the corresponding points in the mesh network are recorded according to each pixel point of the image. The point cloud model within the field of view of the virtual telecentric lens camera is generated through the three-dimensional coordinates. If the corresponding points in multiple mesh networks are projected into the same image coordinates, the point closest to the camera coordinate system is retained (only the point cloud on the surface of the mesh model is sampled); the pixel size and image width and height can be adjusted to adjust the size of the generated local model point cloud. In addition, the local model of the weld is cropped based on the predicted weld length information, and only the important parts such as the weld and the surfaces on both sides of the weld are retained to reduce the interference of subsequent point cloud registration.

[0068] Step 6: Get the position of the weld start and end points in the actual point cloud in the camera coordinate system. Filter the point cloud collected by the camera in step 4. At the same time, based on the weld coordinate system, the actual point cloud can be transformed to the weld coordinate system in advance. According to information such as the weld length, the actual point cloud is cropped to retain only important weld and surface information. Then, the actual point cloud is transformed back to the camera coordinate system through inverse transformation. Based on the high-precision point cloud registration algorithm, the homogeneous transformation matrix M4 from the point cloud model to the actual point cloud (steel component point cloud collected by the camera) is calculated; through M4, the starting and ending position coordinates of the weld in the steel construction model are transformed to obtain the position of the starting and ending points of the weld in the camera coordinate system in the actual point cloud.

[0069] Step 7: Transform the weld start and end coordinates in the camera coordinate system in step 6 to the robot earth coordinate system, and weld the workpiece with the expert welding process library. Through the hand-eye calibration matrix M1 in step 1, transform the weld start and end coordinates in the camera coordinate system in step 5 to the robot earth coordinate system, and weld the workpiece with the expert welding process library.

[0070] Parts or structures not specifically described in the present invention may adopt existing technologies or existing products and will not be described in detail here.

[0071] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for improving the compatibility of automated welding weld position positioning, characterized in that: The following steps are involved: Step 1: Use a high-precision 3D camera and welding robot to perform hand-eye calibration and establish the pose matrix M1 between the camera coordinate system and the robot's earth coordinate system; Step 2: Obtain the pose matrix M2 of the actual workpiece in the robot's earth coordinate system; Step 3: Establish the local coordinate system of the weld through the weld digital model analysis; Step 4: Obtain actual workpiece point cloud data; Step 5: Generate a registration template for the point cloud model of the linear weld using the steel component digital model; Step 6: Get the positions of the starting point and end point of the weld in the actual point cloud in the camera coordinate system; Step 7: Transform the coordinates of the weld start and end points in the camera coordinate system in step 6 to the robot earth coordinate system, and weld the workpiece using the expert welding process library.

2. A method for improving the compatibility of automated welding seam positioning according to claim 1, characterized in that: In step 2, the step of obtaining the pose matrix M2 of the actual workpiece in the robot's earth coordinate system includes: 21) Touch at least three points on the actual workpiece with the tip of the welding gun tool. The distance between the points should be as evenly distributed as possible on the workpiece to improve the calculation accuracy, and obtain the coordinate value (X, Y, Z) of the current robot tool TCP point; 22) Read the model point coordinate (X, Y, Z) value corresponding to the robot TCP point in step 21) in the 3D model software; 23) Solve the pose matrix M2 through the two sets of corresponding points obtained in step 21) and step 22).

3. The method for improving the compatibility of automated welding seam positioning according to claim 1, characterized in that: In step 2, the step of obtaining the pose matrix M2 of the actual workpiece in the robot's earth coordinate system includes: The entire workpiece is photographed using a large-field line scan camera, and the pose matrix M2 is calculated using the 3D registration algorithm in combination with the workpiece digital model.

4. A method for improving the compatibility of automated welding seam positioning according to claim 1, characterized in that: In step 3, the steps of establishing the local coordinate system of the weld by analyzing the weld digital model include: 31) Analyze the digital model of the steel structure to be welded and obtain the standard position information of the weld; 32) Establish the starting and ending coordinates of the weld; 33) The weld position and the normal vectors of the two side surfaces are transformed to the robot earth coordinate system through the pose matrix M2, and the weld coordinate system is established using the transformed weld position normal vector to represent the position of the weld in the robot earth coordinate system.

5. A method for improving the compatibility of automated welding seam positioning according to claim 4, characterized in that: The welds include flat welds and vertical welds. The principles for judging flat welds and vertical welds are as follows: According to the angle between the weld normal vector and the Z-axis normal vector of the robot's geodetic coordinate system, an angle >70° is a flat weld, and an angle <30° is a vertical weld.

6. A method for improving the compatibility of automated welding seam positioning according to claim 5, characterized in that: The steps to establish the coordinate system at the starting point of the flat weld include: 311) The normal vectors of the two sides of the weld F1 and F2 are known. The two end points of the weld, A and B; Calculate the angle between the normal vectors of faces F1 and F2 and the Z-axis normal vector of the world coordinate system to determine the Z-axis, take the side with the smallest angle, and use the surface normal vector as the basis vector of the Z-axis 312) Use Cross product weld vector or Right now or is the normal vector of the faces standing on both sides of the weld; 313) Use Multiply the normal vector of the other face by the dot to determine the angle. If the angle is less than 10°, then the vector or As the basis vector of the Y axis of the coordinate system; 314) According to the vector Determine the X-axis basis vector; 315) Determine the vector that meets the condition of step 312) or Use point A or point B as the origin of the coordinate system; this calculation method is not effective when the angle between surface F1 and surface F2 is not 90°. The vertical plane of the two faces that are not parallel to the weld; 316) Establish the weld end point coordinate system: Replace the origin of the coordinate system calculated in the first five steps with another point as the weld end point coordinate system.

7. A method for improving the compatibility of automated welding seam positioning according to claim 5, characterized in that: The steps to establish the coordinate system at the starting point of the vertical weld include: 411) The normal vectors of the two sides of the weld F1 and F2 are known. The two end points of the weld, A and B; The vector of the straight line connecting the start and end points of the weld or As the Z-axis basis vector of the coordinate system, determine or Is the normal vector of the world coordinate system in the same direction? If so, it is considered or As the Z-axis basis vector of the weld coordinate system At the same time, point A or point B is identified as the starting point of the weld, otherwise, point B or point A is identified as the end point of the weld; 412) Use the normal vector of face F1 and the normal vector of face F2 Perform a cross product if The vector and If the axes are in the same direction, then As the X-axis basis vector, otherwise As the X-axis basis vector; 413) According to the vector Determine the Y-axis basis vector; 414) The weld start and end coordinates A or B determined in step 411) are used as the origin of the weld start or end coordinate system, and the weld coordinate system is established. If the angle between the surface F1 and the surface F2 is not 90°, then Not parallel to surface F2.

8. The method for improving the compatibility of automated welding seam positioning according to claim 1, characterized in that: In step 4, the steps of obtaining actual workpiece point cloud data include: Through the RRT* trajectory planning algorithm, the welding robot is moved to the vicinity of the actual workpiece; The shooting posture Rx, Ry, and Rz of the high-precision 3D camera are set with an empirical value relative to the weld coordinate system. According to the camera field of view and anti-collision algorithm, the appropriate shooting distance d is calculated. The posture matrix M3 of the camera relative to the weld coordinate system is calculated based on Rx, Ry, Rz, and d. Through the M1, M2, and M3 posture matrix transformations, the welding robot is moved to the weld coordinate system, and a high-precision 3D camera is used to take photos to obtain the actual workpiece point cloud data.

9. A method for improving the compatibility of automated welding seam positioning according to claim 1, characterized in that: In step 5, the steps of using the steel component digital model to generate a registration template of the point cloud model of the linear weld include: 51) The weld coordinate system is obtained from step 2, and the camera's field of view is known through factory parameters; 52) The workpiece digital model is transformed into the camera coordinate system through the pose matrices M2 and M3; 53) Define a virtual telecentric lens camera internal parameter; 54) According to the camera internal parameters, the digital-analog mesh network coordinates within the field of view of the virtual telecentric lens camera are projected into the image coordinate system. At the same time, the three-dimensional coordinates of the corresponding points in the mesh network are recorded according to each pixel point of the image. The point cloud model within the field of view of the virtual telecentric lens camera is generated by the three-dimensional coordinates. If the corresponding points in multiple mesh networks are projected into the same image coordinates, the points closest to the camera coordinate system are retained; the size of the generated local model point cloud can be adjusted by adjusting the pixel size and the image width and height; the local model of the weld is cropped by using the predicted weld length information, and only the important parts of the weld and the surfaces on both sides of the weld are retained to reduce the interference of subsequent point cloud alignment.

10. The method for improving the compatibility of automated welding seam positioning according to claim 1, characterized in that: In step 6, the steps of obtaining the positions of the start and end points of the weld in the actual point cloud in the camera coordinate system include: The point cloud collected by the camera in step 4 is filtered, and based on the weld coordinate system, the actual point cloud is pre-transformed to the weld coordinate system. According to the weld length information, the actual point cloud is cropped to retain only important welds and surface information. Then, the actual point cloud is transformed back to the camera coordinate system through inverse transformation. Based on the high-precision point cloud registration algorithm, the homogeneous transformation matrix M4 from the point cloud model to the actual point cloud or the steel component point cloud collected by the camera is calculated, thereby transforming the starting and ending position coordinates of the weld in the steel construction model to obtain the positions of the starting and ending points of the weld in the actual point cloud in the camera coordinate system.

Images