Robot auto-induction brazing equipment based on visual guidance and method thereof

Abstract

The invention relates to robot automatic induction brazing equipment and method based on visual guidance, an induction brazing welding gun is fixedly connected to a robot, and the induction brazing welding gun and a 3D camera face a welding point area of a circular tube object; a conversion relation T (R, t) between a camera coordinate system and a robot coordinate system is obtained through a hand-eye calibration method, and the camera coordinate system is unified to the robot coordinate system; the camera photographs a welding point area of the circular tube object and collects an image of the welding point area; precisely positioning a welding point of the circular tube object by utilizing a visual algorithm to obtain a pose P (x, y, z, Rx, Ry and Rz) of the welding point in the camera; and according to the coordinate system conversion relation T (R, t) determined through hand-eye calibration, the welding point pose P (x, y, z, Rx, Ry and Rz) is converted from a camera coordinate system to a robot coordinate system, and a robot drives an induction brazing welding gun to conduct welding. The object is accurately positioned through photographing of the 3D camera, and the robot is guided to achieve high-precision automatic induction brazing according to the posture of the object.

Description

Technical Field

[0001] The invention relates to robot automatic induction brazing equipment based on vision guidance and a method thereof. Background Art

[0002] The traditional welding process is to use flame heating to melt the welding wire for welding. The welding process will produce a lot of smoke and dust, and the heating efficiency is low. Induction brazing is a new welding process that can complete welding efficiently through electromagnetic induction heating. It is green and environmentally friendly and has gradually replaced flame welding.

[0003] At present, induction brazing is performed by manually operating induction brazing equipment, which requires a handheld welding gun to move the two induction blocks of the welding gun to the two sides of the round tube object to ensure that the object is evenly heated at the center of the two induction blocks. In the height direction, the joint of the pipe to be welded (the interface of the two cylinders) must be located at the center height of the induction block, and the round tube must be parallel to the inner wall of the induction block. At the same time, the welding wire needs to be aligned with the welding point on the object. Due to the obstruction of vision, it is difficult to achieve accurate wire feeding by manual vision. At the same time, due to the poor stability of the handheld welding gun operation, it will lead to poor welding quality and high re-welding rate, which seriously reduces production efficiency.

[0004] Therefore, it is necessary to develop equipment and methods for automatic induction brazing. Summary of the invention

[0005] The purpose of the present invention is to overcome the deficiencies in the prior art and to provide a robot automatic induction brazing device and method based on vision guidance, aiming to solve the problems of low efficiency and poor quality of manual welding.

[0006] The purpose of the present invention is achieved through the following technical solutions:

[0007] The robot automatic induction brazing equipment based on vision guidance is characterized by: comprising a 3D camera for collecting images of the welding point area of ​​a round tube object, an induction brazing gun for automatic induction brazing of the round tube object, and a robot for carrying the induction brazing gun for movement, the induction brazing gun is fixedly connected to the robot, and the induction brazing gun and the 3D camera are facing the welding point area of ​​the round tube object.

[0008] Furthermore, in the above-mentioned vision-guided robot automatic induction brazing equipment, the 3D camera is fixedly connected to the end of the robot, or installed next to the round tube object, and is aimed at the welding point area of ​​the round tube object.

[0009] Furthermore, in the above-mentioned vision-guided robotic automatic induction brazing equipment, the 3D camera is a line laser profile camera or a photographic structured light camera.

[0010] Furthermore, in the above-mentioned vision-guided robot automatic induction brazing equipment, the robot is a six-axis robot, a four-axis robot or an XYZ-axis motion module.

[0011] The present invention provides a robot automatic induction brazing method based on vision guidance, comprising the following steps:

[0012] (1) Obtain the transformation relationship T(R, t) between the camera coordinate system and the robot coordinate system through the hand-eye calibration method, and convert the camera coordinate system to the robot coordinate system, where the hand is defined as the coordinate system of the robot end and the eye is defined as the coordinate system of the camera;

[0013] (2) The camera takes a picture of the welding point area of ​​the circular tube object to collect an image of the welding point area;

[0014] (3) Use the visual algorithm to accurately locate the welding point of the circular tube object and obtain the position P(x, y, z, Rx, Ry, Rz) of the welding point in the camera;

[0015] (4) According to the coordinate system conversion relationship T(R, t) determined by hand-eye calibration, the welding point posture P(x, y, z, Rx, Ry, Rz) is converted from the camera coordinate system to the robot coordinate system, and the robot drives the induction brazing gun to perform welding.

[0016] Furthermore, in the above-mentioned vision-guided robot automatic induction brazing method, the 3D camera is fixedly connected to the end of the robot, and the steps of the hand-eye calibration method are:

[0017] 11a) Prepare a finely machined calibration ball, place it in the robot's workspace and keep it stationary relative to the robot's base, and let the position of the ball center relative to the robot's base be Y(t);

[0018] 11b) Let the robot take pictures of the calibration ball with a camera in different postures to obtain its point cloud. Let the i-th posture be A i (R, t);

[0019] 11c) Perform spherical fitting on the obtained point cloud to obtain the position of the sphere center in the camera coordinate system, and let the i-th sphere center be B i (t);

[0020] 11d) Let the hand-eye calibration result to be calculated be X(R, t), then after collecting n postures, the following matrix equations are obtained:

[0021]

[0022] 11e) Optimize and solve it to obtain the hand-eye calibration result X(R, t), which is the transformation relationship T(R, t) between the camera coordinate system and the robot coordinate system.

[0023] Furthermore, in the above-mentioned vision-guided robot automatic induction brazing method, the 3D camera is installed next to the round tube object, and the steps of the hand-eye calibration method are as follows:

[0024] 12a) Prepare a precision-machined calibration ball and fix it on the end flange of the robot. Let the position of the ball center relative to the end flange of the robot be Y(t);

[0025] 12b) Let the robot face the camera with the marker ball in different postures. The camera takes pictures of the calibration ball and obtains its point cloud. Let the i-th posture be A i (R, t);

[0026] 12c) Perform spherical fitting on the obtained point cloud to obtain the position of the sphere center in the camera coordinate system, and let the i-th sphere center be B i (t);

[0027] 12d) Let the hand-eye calibration result to be calculated be X(R, t), then after collecting n postures, the following matrix equations are obtained:

[0028]

[0029] 12e) Optimize and solve it to obtain the hand-eye calibration result X(R, t), which is the transformation relationship T(R, t) between the camera coordinate system and the robot coordinate system.

[0030] Furthermore, in the above-mentioned robot automatic induction brazing method based on vision guidance, the welding point of the round tube object is visually positioned in the following steps:

[0031] First, the point cloud image collected by the 3D camera is preprocessed to reduce noise and background.

[0032] Then, the RANSAC algorithm is used to fit two cylinders with different diameters. For a cylinder with a given radius, the RANSAC fitting method is:

[0033] Estimate the normal vector of each point using several points in its neighborhood;

[0034] Randomly sample all points in the point cloud, and use the coordinates and normal vectors of the sampled points to perform cylinder fitting using the least squares method to obtain the cylinder parameters;

[0035] If the fitting error is large, the cylinder fitting is performed again as described above; if the fitting error is small, further calculation is performed on which points in the point cloud are also on the fitted cylindrical model, and for all points on the cylindrical model, the cylinder fitting is continued using the least squares method;

[0036] Repeat until the number of iterations reaches the threshold N, or the difference between the result of the next iteration and the previous iteration is less than the threshold σ;

[0037] Finally, calculate the position P(x, y, z, Rx, Ry, Rz) of the welding point in the camera, where (x, y, z) is the center of the cylinder at the intersection of the two cylinders, and (Rx, Ry, Rz) is the direction of the central axis of the cylinder; the calculation method is:

[0038] Project all points on the surface of the large cylinder onto the central axis of the large cylinder; project all points on the surface of the small cylinder onto the central axis of the small cylinder;

[0039] For all points projected onto the central axis of the large cylinder, take the point closest to the small cylinder; for all points projected onto the central axis of the small cylinder, take the point closest to the large cylinder;

[0040] For the two points obtained, the midpoint is defined as the center of the cylinder where the two cylinders meet (x, y, z), and the axis of the large cylinder or the small cylinder is defined as the direction of the cylinder centerline (Rx, Ry, Rz), and the position P (x, y, z, Rx, Ry, Rz) of the welding point in the camera is obtained.

[0041] Furthermore, in the above-mentioned robot automatic induction brazing method based on vision guidance, the welding point of the round tube object is visually positioned in the following steps:

[0042] First, the point cloud image collected by the 3D camera is preprocessed to reduce noise and background.

[0043] Then, the RANSAC algorithm is used to fit two cylinders with different diameters. For a cylinder with a given radius, the RANSAC fitting method is:

[0044] Estimate the normal vector of each point using several points in its neighborhood;

[0045] Randomly sample all points in the point cloud, and use the coordinates and normal vectors of the sampled points to perform cylinder fitting using the least squares method to obtain the cylinder parameters;

[0046] If the fitting error is large, the cylinder fitting is performed again as described above; if the fitting error is small, further calculation is performed on which points in the point cloud are also on the fitted cylindrical model, and for all points on the cylindrical model, the cylinder fitting is continued using the least squares method;

[0047] Repeat until the number of iterations reaches the threshold N, or the difference between the result of the next iteration and the previous iteration is less than the threshold σ;

[0048] Finally, the visual positioning method of the welding point of the circular tube object, after obtaining the two large and small cylinders, generates a stepped standard point cloud template containing the two cylinders according to the radius of the two cylinders. The center point of the intersection of the two cylinders and the direction of the central axis of the cylinder are known in the template; template matching is performed in the point cloud obtained by the 3D camera to obtain the position P(x, y, z, Rx, Ry, Rz) of the welding point in the camera.

[0049] Furthermore, in the above-mentioned robot automatic induction brazing method based on vision guidance, the welding point of the round tube object is visually positioned in the following steps:

[0050] First, the point cloud image collected by the 3D camera is preprocessed to reduce noise and background.

[0051] Then, find the edge in the point cloud, use the points on the edge to fit the circle, get the center point (x, y, z) of the fitted circle and the normal direction of the center point (Rx, Ry, Rz), and get the position P (x, y, z, Rx, Ry, Rz) of the welding point in the camera.

[0052] Furthermore, in the above-mentioned robot automatic induction brazing method based on vision guidance, the welding point of the round tube object is visually positioned in the following steps:

[0053] First, offline teaching is performed to collect the 3D point cloud Points (x, y, z) of the round tube object used as the teaching welding point, and to teach the robot welding reference pose Robot (x, y, z, Rx, Ry, Rz);

[0054] Then, during measurement, the 3D point cloud Points (x', y', z') of the circular tube object is collected;

[0055] Then, the 3D point cloud matching method is used to calculate the pose offset (△x, △y, △z, △Rx, △Ry, △Rz) between the offline object 3D point cloud and the online object 3D point cloud;

[0056] Finally, the coordinate system transformation relationship T(R, t) determined by hand-eye calibration is used to transform the pose offset P(△x,△y,△z,△Rx,△Ry,△Rz) from the camera coordinate system to the robot coordinate system. The converted offset is added to the robot reference pose to obtain Robot(x+△x,y+△y,z+△z,Rx+△Rx,Ry+△Ry,Rz+△Rz). The robot drives the induction brazing gun for welding.

[0057] Compared with the prior art, the present invention has significant advantages and beneficial effects, which are specifically embodied in the following aspects:

[0058] The present invention provides a robot induction brazing device and method for round tube objects with visual guidance. The object is accurately positioned by taking photos with a 3D camera, and the robot is guided to achieve high-precision automatic induction brazing according to the object's posture. High-precision visual positioning is used to guide a high-precision, high-repeatability robot to complete automatic welding. The camera takes photos to achieve high-precision positioning of the object, which solves the problem of inaccurate manual vision. Robot welding solves the problem of instability in the operation of a handheld welding gun, achieving high-quality and high-efficiency welding results.

[0059] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by practicing the specific embodiments of the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0061] Figure 1 : A schematic diagram of an installation form of a 3D camera of the device of the present invention;

[0062] Figure 2 : A schematic diagram of another installation form of the 3D camera of the device of the present invention. DETAILED DESCRIPTION

[0063] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.

[0064] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, directional terms and order terms are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0065] The robot automatic induction brazing equipment based on vision guidance comprises a 3D camera 103 for collecting images of welding point areas of round tube objects, an induction brazing welding gun 102 for automatic induction brazing of round tube objects, and a robot 101 for carrying the induction brazing welding gun 102 for movement. The induction brazing welding gun 102 is fixedly connected to the robot 101. Figure 1 , the 3D camera 103 is fixedly connected to the end of the robot 101, or, as Figure 2 , the 3D camera 103 is installed next to the round tube object, and the induction brazing welding gun 102 and the 3D camera 103 are directed toward the welding point area 104 of the round tube object.

[0066] The 3D camera 103 is a line laser profile camera or a photographic structured light camera. The robot 101 is a six-axis robot, a four-axis robot or an XYZ axis motion module.

[0067] The present invention provides a robot automatic induction brazing method based on vision guidance, comprising the following steps:

[0068] (1) Obtain the transformation relationship T(R, t) between the camera coordinate system and the robot coordinate system through the hand-eye calibration method, and convert the camera coordinate system to the robot coordinate system, where the hand is defined as the coordinate system of the robot end and the eye is defined as the coordinate system of the camera;

[0069] (2) The camera takes a picture of the welding point area of ​​the circular tube object to collect an image of the welding point area;

[0070] (3) Use the visual algorithm to accurately locate the welding point of the circular tube object and obtain the position P (x, y, z, Rx, Ry, Rz) of the welding point in the camera;

[0071] (4) According to the coordinate system conversion relationship T(R, t) determined by hand-eye calibration, the welding point posture P(x, y, z, Rx, Ry, Rz) is converted from the camera coordinate system to the robot coordinate system, and the robot drives the induction brazing gun to perform welding.

[0072] Based on the 3D camera fixedly connected to the end of the robot, the steps of the hand-eye calibration method are:

[0073] 11a) Prepare a finely machined calibration ball, place it in the robot's workspace and keep it stationary relative to the robot's base, and let the position of the ball center relative to the robot's base be Y(t);

[0074] 11b) Let the robot take pictures of the calibration ball with a camera in different postures to obtain its point cloud. Let the i-th posture be A i (R, t);

[0075] 11c) Perform spherical fitting on the obtained point cloud to obtain the position of the sphere center in the camera coordinate system, and let the i-th sphere center be B i (t);

[0076] 11d) Let the hand-eye calibration result to be calculated be X(R, t), then after collecting n postures, the following matrix equations are obtained:

[0077]

[0078] 11e) Optimize and solve it to obtain the hand-eye calibration result X(R, t), which is the transformation relationship T(R, t) between the camera coordinate system and the robot coordinate system.

[0079] Based on the 3D camera installed next to the round tube object, the steps of the hand-eye calibration method are:

[0080] 12a) Prepare a precision-machined calibration ball and fix it on the end flange of the robot. Let the position of the ball center relative to the end flange of the robot be Y(t);

[0081] 12b) Let the robot face the camera with the marker ball in different postures. The camera takes pictures of the calibration ball and obtains its point cloud. Let the i-th posture be A i (R, t);

[0082] 12c) Perform spherical fitting on the obtained point cloud to obtain the position of the sphere center in the camera coordinate system, and let the i-th sphere center be B i (t);

[0083] 12d) Let the hand-eye calibration result to be calculated be X(R, t), then after collecting n postures, the following matrix equations are obtained:

[0084]

[0085] 12e) Optimize and solve it to obtain the hand-eye calibration result X(R, t), which is the transformation relationship T(R, t) between the camera coordinate system and the robot coordinate system.

[0086] One of the methods for visual positioning of welding points of round tube objects is:

[0087] First, the point cloud image collected by the 3D camera is preprocessed to reduce noise and background.

[0088] Then, the RANSAC algorithm is used to fit two cylinders with different diameters. For a cylinder with a given radius, the RANSAC fitting method is:

[0089] Estimate the normal vector of each point using several points in its neighborhood;

[0090] Randomly sample all points in the point cloud, and use the coordinates and normal vectors of the sampled points to perform cylinder fitting using the least squares method to obtain the cylinder parameters;

[0091] If the fitting error is large, the cylinder fitting is performed again as described above; if the fitting error is small, further calculation is performed on which points in the point cloud are also on the fitted cylindrical model, and for all points on the cylindrical model, the cylinder fitting is continued using the least squares method;

[0092] Repeat until the number of iterations reaches the threshold N, or the difference between the result of the next iteration and the previous iteration is less than the threshold σ;

[0093] Finally, calculate the position P(x, y, z, Rx, Ry, Rz) of the welding point in the camera, where (x, y, z) is the center of the cylinder at the intersection of the two cylinders, and (Rx, Ry, Rz) is the direction of the central axis of the cylinder; the calculation method is:

[0094] Project all points on the surface of the large cylinder onto the central axis of the large cylinder; project all points on the surface of the small cylinder onto the central axis of the small cylinder;

[0095] For all points projected onto the central axis of the large cylinder, take the point closest to the small cylinder; for all points projected onto the central axis of the small cylinder, take the point closest to the large cylinder;

[0096] For the two points obtained, the midpoint is defined as the center of the cylinder where the two cylinders meet (x, y, z), and the axis of the large cylinder or the small cylinder is defined as the direction of the cylinder centerline (Rx, Ry, Rz), and the position P (x, y, z, Rx, Ry, Rz) of the welding point in the camera is obtained.

[0097] The second method for visual positioning of welding points of round tube objects is:

[0098] First, the point cloud image collected by the 3D camera is preprocessed to reduce noise and background.

[0099] Then, the RANSAC algorithm is used to fit two cylinders with different diameters. For a cylinder with a given radius, the RANSAC fitting method is:

[0100] Estimate the normal vector of each point using several points in its neighborhood;

[0101] Randomly sample all points in the point cloud, and use the coordinates and normal vectors of the sampled points to perform cylinder fitting using the least squares method to obtain the cylinder parameters;

[0102] If the fitting error is large, the cylinder fitting is performed again as described above; if the fitting error is small, further calculation is performed on which points in the point cloud are also on the fitted cylindrical model, and for all points on the cylindrical model, the cylinder fitting is continued using the least squares method;

[0103] Repeat until the number of iterations reaches the threshold N, or the difference between the result of the next iteration and the previous iteration is less than the threshold σ;

[0104] Finally, the visual positioning method of the welding point of the circular tube object, after obtaining the two large and small cylinders, generates a stepped standard point cloud template containing the two cylinders according to the radius of the two cylinders. The center point of the intersection of the two cylinders and the direction of the central axis of the cylinder are known in the template; template matching is performed in the point cloud obtained by the 3D camera to obtain the position P(x, y, z, Rx, Ry, Rz) of the welding point in the camera.

[0105] For the visual positioning of the welding points of round tube objects, the third method is:

[0106] First, the point cloud image collected by the 3D camera is preprocessed to reduce noise and background.

[0107] Then, find the edge in the point cloud, use the points on the edge to fit the circle, get the center point (x, y, z) of the fitted circle and the normal direction of the center point (Rx, Ry, Rz), and get the position P (x, y, z, Rx, Ry, Rz) of the welding point in the camera.

[0108] For the visual positioning of the welding points of round tube objects, the fourth method is:

[0109] First, offline teaching is performed to collect the 3D point cloud Points (x, y, z) of the round tube object used as the teaching welding point, and to teach the robot welding reference pose Robot (x, y, z, Rx, Ry, Rz);

[0110] Then, during measurement, the 3D point cloud Points (x', y', z') of the circular tube object is collected;

[0111] Then, the 3D point cloud matching method is used to calculate the pose offset (△x, △y, △z, △Rx, △Ry, △Rz) between the offline object 3D point cloud and the online object 3D point cloud;

[0112] Finally, the coordinate system transformation relationship T(R, t) determined by hand-eye calibration is used to transform the pose offset P(△x,△y,△z,△Rx,△Ry,△Rz) from the camera coordinate system to the robot coordinate system. The converted offset is added to the robot reference pose to obtain Robot(x+△x,y+△y,z+△z,Rx+△Rx,Ry+△Ry,Rz+△Rz). The robot drives the induction brazing gun for welding.

[0113] In summary, the present invention provides a robot induction brazing device and method for visually guided round tube objects. By taking pictures with a 3D camera, the object is accurately positioned, and the robot is guided to achieve high-precision automatic induction brazing according to the object's posture. High-precision visual positioning is used to guide a high-precision, high-repeatability robot to complete automatic welding. The camera takes pictures to achieve high-precision positioning of the object, which solves the problem of inaccurate manual vision. Robot welding solves the problem of instability in the operation of a handheld welding gun, achieving high-quality and high-efficiency welding results.

[0114] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention. It should be noted that similar numbers and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings.

[0115] The above is only a specific implementation of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

[0116] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

Claims

1. Robotic automatic induction brazing equipment based on vision guidance, characterized by: The invention comprises a 3D camera (103) for collecting images of a welding point area of ​​a circular tube object, an induction brazing welding gun (102) for automatically induction brazing the circular tube object, and a robot (101) for carrying the induction brazing welding gun (102) for movement. The induction brazing welding gun (102) is fixedly connected to the robot (101), and the induction brazing welding gun (102) and the 3D camera (103) face the welding point area of ​​the circular tube object.

2. The vision-guided robot automatic induction brazing equipment according to claim 1 is characterized in that: The 3D camera (103) is fixedly connected to the end of the robot (101), or installed beside the round tube object, and is aimed at the welding point area of ​​the round tube object.

3. The vision-guided robot automatic induction brazing equipment according to claim 1 or 2, characterized in that: The 3D camera (103) is a line laser profile camera or a photographic structured light camera.

4. The vision-guided robot automatic induction brazing equipment according to claim 1 is characterized in that: The robot (101) is a six-axis robot, a four-axis robot or an XYZ-axis motion module.

5. A method for realizing a robot automatic induction brazing method based on vision guidance using the device of claim 1, characterized in that: The following steps are involved: (1) Obtain the transformation relationship T(R, t) between the camera coordinate system and the robot coordinate system through the hand-eye calibration method, and convert the camera coordinate system to the robot coordinate system, where the hand is defined as the coordinate system of the robot end and the eye is defined as the coordinate system of the camera; (2) The camera takes a picture of the welding point area of ​​the circular tube object to collect an image of the welding point area; (3) Use the visual algorithm to accurately locate the welding point of the circular tube object and obtain the position P(x, y, z, Rx, Ry, Rz) of the welding point in the camera; (4) According to the coordinate system conversion relationship T(R, t) determined by hand-eye calibration, the welding point posture P(x, y, z, Rx, Ry, Rz) is converted from the camera coordinate system to the robot coordinate system, and the robot drives the induction brazing gun to perform welding.

6. The method of automatic induction brazing by robot based on vision guidance according to claim 5, characterized in that: The 3D camera is fixedly connected to the end of the robot. The steps of the hand-eye calibration method are as follows: 11a) Prepare a finely machined calibration ball, place it in the robot's workspace and keep it stationary relative to the robot's base, and let the position of the ball center relative to the robot's base be γ(t); 11b) Let the robot take pictures of the calibration ball with a camera in different postures to obtain its point cloud. Let the i-th posture be A i (R, t); 11c) Perform spherical fitting on the obtained point cloud to obtain the position of the sphere center in the camera coordinate system, and let the i-th sphere center be B i (t); 11d) Let the hand-eye calibration result to be calculated be X(R, t), then after collecting n postures, the following matrix equations are obtained: 11e) Optimize and solve it to obtain the hand-eye calibration result X(R, t), which is the transformation relationship T(R, t) between the camera coordinate system and the robot coordinate system.

7. The method of automatic induction brazing by robot based on vision guidance according to claim 5, characterized in that: The 3D camera is installed next to the round tube object, and the steps of the hand-eye calibration method are as follows: 12a) Prepare a precision-machined calibration ball and fix it on the end flange of the robot. Let the position of the ball center relative to the end flange of the robot be Y(y); 12b) Let the robot face the camera with the marker ball in different postures. The camera takes pictures of the calibration ball and obtains its point cloud. Let the i-th posture be A i (R, t); 12c) Perform spherical fitting on the obtained point cloud to obtain the position of the sphere center in the camera coordinate system, and let the i-th sphere center be B i (t); 12d) Let the hand-eye calibration result to be calculated be X(R, t), then after collecting n postures, the following matrix equations are obtained: 12e) Optimize and solve it to obtain the hand-eye calibration result X(R, t), which is the transformation relationship T(R, t) between the camera coordinate system and the robot coordinate system.

8. The method of automatic induction brazing by robot based on vision guidance according to claim 5, characterized in that: Visual positioning of the welding point of a round tube object, the steps are: First, the point cloud image collected by the 3D camera is preprocessed to reduce noise and background. Then, the RANSAC algorithm is used to fit two cylinders with different diameters. For a cylinder with a given radius, the RANSAC fitting method is: Estimate the normal vector of each point using several points in its neighborhood; Randomly sample all points in the point cloud, and use the coordinates and normal vectors of the sampled points to perform cylinder fitting using the least squares method to obtain the cylinder parameters; If the fitting error is large, the cylinder fitting is performed again as described above; if the fitting error is small, further calculation is performed on which points in the point cloud are also on the fitted cylindrical model, and for all points on the cylindrical model, the cylinder fitting is continued using the least squares method; Repeat until the number of iterations reaches the threshold N, or the difference between the result of the next iteration and the previous iteration is less than the threshold σ; Finally, calculate the position P(x, y, z, Rx, Ry, Rz) of the welding point in the camera, where (x, y, z) is the center of the cylinder at the intersection of the two cylinders, and (Rx, Ry, Rz) is the direction of the central axis of the cylinder; the calculation method is: Project all points on the surface of the large cylinder onto the central axis of the large cylinder; project all points on the surface of the small cylinder onto the central axis of the small cylinder; For all points projected onto the central axis of the large cylinder, take the point closest to the small cylinder; for all points projected onto the central axis of the small cylinder, take the point closest to the large cylinder; For the two points obtained, the midpoint is defined as the center of the cylinder where the two cylinders meet (x, y, z), and the axis of the large cylinder or the small cylinder is defined as the direction of the cylinder centerline (Rx, Ry, Rz), and the position P (x, y, z, Rx, Ry, Rz) of the welding point in the camera is obtained.

9. The method of automatic induction brazing by robot based on vision guidance according to claim 5, characterized in that: Visual positioning of the welding point of a round tube object, the steps are: First, the point cloud image collected by the 3D camera is preprocessed to reduce noise and background. Then, the RANSAC algorithm is used to fit two cylinders with different diameters. For a cylinder with a given radius, the RANSAC fitting method is: Estimate the normal vector of each point using several points in its neighborhood; Randomly sample all points in the point cloud, and use the coordinates and normal vectors of the sampled points to perform cylinder fitting using the least squares method to obtain the cylinder parameters; If the fitting error is large, the cylinder fitting is performed again as described above; if the fitting error is small, further calculation is performed on which points in the point cloud are also on the fitted cylindrical model, and for all points on the cylindrical model, the cylinder fitting is continued using the least squares method; Repeat until the number of iterations reaches the threshold N, or the difference between the result of the next iteration and the previous iteration is less than the threshold σ; Finally, the visual positioning method of the welding point of the circular tube object, after obtaining the two large and small cylinders, generates a stepped standard point cloud template containing the two cylinders according to the radius of the two cylinders. The center point of the intersection of the two cylinders and the direction of the central axis of the cylinder are known in the template; template matching is performed in the point cloud obtained by the 3D camera to obtain the position P(x, y, z, Rx, Ry, Rz) of the welding point in the camera.

10. The method for automatic induction brazing by a robot based on vision guidance according to claim 5, characterized in that: Visual positioning of the welding point of a round tube object, the steps are: First, the point cloud image collected by the 3D camera is preprocessed to reduce noise and background. Then, find the edge in the point cloud, use the points on the edge to fit the circle, get the center point (x, y, z) of the fitted circle and the normal direction of the center point (Rx, Ry, Rz), and get the position P (x, y, z, Rx, Ry, Rz) of the welding point in the camera.

11. The method for automatic induction brazing by a robot based on vision guidance according to claim 5, characterized in that: Visual positioning of the welding point of a round tube object, the steps are: First, offline teaching is performed to collect the 3D point cloud Points (x, y, z) of the round tube object used as the teaching welding point, and to teach the robot welding reference pose Robot (x, y, z, Rx, Ry, Rz); Then, during measurement, the 3D point cloud Points (x', y', z') of the circular tube object is collected; Then, the 3D point cloud matching method is used to calculate the pose offset (△x, △y, △z, △Rx, △Ry, △Rz) between the offline object 3D point cloud and the online object 3D point cloud; The coordinate system conversion relationship T(R, t) determined by hand-eye calibration is used to convert the pose offset P(△x,△y,△z,△Rx,△Ry,△Rz) from the camera coordinate system to the robot coordinate system. The converted offset is added to the robot reference pose to obtain Robot(x+△x,y+△y,z+△z,Rx+△Rx,Ry+△Ry,Rz+△Rz). The robot drives the induction brazing gun for welding.

Images