A teach-free welding method and system based on laser vision sensors

Abstract

This invention discloses a teach-free welding method and system based on a laser vision sensor. The method includes: calibrating and obtaining the rotation center and normal vector of the external axis relative to the robot's rotation plane; calibrating and obtaining the conversion relationship between the robot's end effector and the laser vision sensor; calculating and generating a theoretical welding trajectory model of the box based on key feature parameters of the box, where the key feature parameters include at least the side length of the box; performing trajectory planning based on the theoretical welding trajectory model and the rotation angle of the external axis to generate a set of trajectories for visual scanning by the laser vision sensor; and calculating the welding trajectory based on the trajectory set and the conversion relationship, so that the robot and the external axis can perform linked welding on the box. This invention can reduce the difficulty of welding operations, lower the threshold for using robotic arms for welding, and achieve high welding efficiency and high weld consistency.

Description

Technical Field

[0001] This invention relates to the field of welding technology, and in particular to a teach-free welding method and system based on a laser vision sensor. Background Technology

[0002] Oil-immersed transformers are a new type of high-performance transformer with a more rational structure and superior performance. Currently, the welding of oil-immersed transformer tanks is mostly done manually, but manual welding faces challenges such as difficulty in recruiting welders, and the need for overhead cranes to assist in the rotation of large tanks. Some systems use welding robots in conjunction with vision systems for automated welding, but these generally suffer from complex teaching methods for vision imaging positions. A single set of teaching positions is only applicable to the welding of transformer tanks of the same model, and developing teaching imaging positions for welding new tanks is time-consuming and complex to program. Summary of the Invention

[0003] In view of the above technical problems, the present invention provides a teach-free welding method and system based on a laser vision sensor to solve the problems of difficult and complex welding of oil-immersed transformer tanks in the prior art.

[0004] Other features and advantages of the invention will become apparent from the following detailed description, or may be learned in part by practice of the invention.

[0005] According to one aspect of the present invention, a teachless welding method based on a laser vision sensor is provided, the method comprising:

[0006] The calibration obtains the normal vector of the robot's external axis relative to the robot's rotation center and rotation plane to determine the relationship between the external axis and the robot. The external axis is used to drive the box to rotate in one or more directions, and the box has a regular shape.

[0007] The conversion relationship between the robot end effector and the laser vision sensor is calibrated and obtained. The laser vision sensor is used to obtain the vision of the box located on the external axis, and the conversion relationship between it and the robot end effector is fixed.

[0008] Based on the key feature parameters of the box, a theoretical welding trajectory model of the box is calculated and generated. The key feature parameters include at least the side length of the box.

[0009] Based on the theoretical welding trajectory model and the rotation angle of the external shaft, trajectory planning is performed to generate a set of trajectories for visual scanning by the laser vision sensor;

[0010] Based on the trajectory set and the transformation relationship, the welding trajectory is calculated when the laser vision sensor scans the weld seam of the box, so that the robot and the external axis can perform linkage welding on the box.

[0011] Furthermore, the calibration process for obtaining the normal vectors of the robot's external axes relative to the robot's rotation center and rotation plane specifically includes:

[0012] A calibrated marker is set on the mounting plate plane of the outer shaft for mounting the housing, and the outer shaft is driven to rotate at multiple angles;

[0013] The coordinates of the robot end effector pointing to the fixed point marker at different angles are obtained, and combined with the rotation angle, the rotation center and normal vector of the mounting plate are obtained.

[0014] Furthermore, the calibration process for obtaining the conversion relationship between the robot's end effector and the laser vision sensor specifically includes:

[0015] The laser vision sensor is controlled to take pictures of a fixed calibration plate in multiple different postures, and a homogeneous transformation equation is established based on the hand-eye relationship.

[0016] The transformation relationship is obtained by solving the homogeneous transformation equation.

[0017] Furthermore, the calculation and generation of the theoretical welding trajectory model of the box based on the key feature parameters of the box includes:

[0018] The side length of the box is obtained. When the rotation axis of the external axis coincides with the line connecting the centroid of the top surface and the centroid of the bottom surface of the box, the distance from the centroid of the top surface of the box to each side of its surface is calculated to obtain the position coordinates of each vertex of the box relative to the coordinate system of the robot. Multiple position coordinates are combined to form the theoretical welding trajectory model.

[0019] Furthermore, the step of planning the trajectory based on the theoretical welding trajectory model and the rotation angle of the external axis to generate a set of trajectories for visual scanning by the laser vision sensor includes:

[0020] When the laser vision sensor performs visual scanning, the external axis is rotated to generate visual search points for each side of the box at different angles. The visual search points are calculated by the theoretical welding trajectory model through the transformation relationship and the fixed offset distance of the robot. The set of visual search points for each side is used as the trajectory set.

[0021] Based on another aspect of the present invention, a teachless welding system based on a laser vision sensor is also provided, the system comprising:

[0022] A robot having multiple degrees of freedom, with a welding torch mounted at its end effector;

[0023] An external shaft is used to fix the box and drive the box to rotate under the robot. The fixed box has a regular shape.

[0024] A laser vision sensor, which is mounted on the end effector of the robot;

[0025] A controller, the controller being used to control the operation of the robot, the external axis and the laser vision sensor, and to perform the following methods during control:

[0026] The calibration obtains the normal vector of the external axis relative to the rotation center and rotation plane of the robot;

[0027] The conversion relationship between the robot end effector and the laser vision sensor is calibrated and obtained;

[0028] Based on the key feature parameters of the box, a theoretical welding trajectory model of the box is calculated and generated. The key feature parameters include at least the side length of the box.

[0029] Based on the theoretical welding trajectory model and the rotation angle of the external shaft, trajectory planning is performed to generate a set of trajectories for visual scanning by the laser vision sensor;

[0030] Based on the trajectory set and the transformation relationship, the welding trajectory is calculated when the laser vision sensor scans the weld seam of the box, so that the robot and the external axis can perform linkage welding on the box.

[0031] Furthermore, the calibration process for obtaining the normal vectors of the robot's external axes relative to the robot's rotation center and rotation plane specifically includes:

[0032] A calibrated marker is set on the mounting plate plane of the outer shaft for mounting the housing, and the outer shaft is driven to rotate at multiple angles;

[0033] The coordinates of the robot end effector pointing to the fixed point marker at different angles are obtained, and combined with the rotation angle, the rotation center and normal vector of the mounting plate are obtained.

[0034] Furthermore, the calibration process for obtaining the conversion relationship between the robot's end effector and the laser vision sensor specifically includes:

[0035] The laser vision sensor is controlled to take pictures of a fixed calibration plate in multiple different postures, and a homogeneous transformation equation is established based on the hand-eye relationship.

[0036] The transformation relationship is obtained by solving the homogeneous transformation equation.

[0037] Furthermore, the calculation and generation of the theoretical welding trajectory model of the box based on the key feature parameters of the box includes:

[0038] The side length of the box is obtained. When the rotation axis of the external axis coincides with the line connecting the centroid of the top surface and the centroid of the bottom surface of the box, the distance from the centroid of the top surface of the box to each side of its surface is calculated to obtain the position coordinates of each vertex of the box relative to the coordinate system of the robot. Multiple position coordinates are combined to form the theoretical welding trajectory model.

[0039] Furthermore, the step of planning the trajectory based on the theoretical welding trajectory model and the rotation angle of the external axis to generate a set of trajectories for visual scanning by the laser vision sensor includes:

[0040] When the laser vision sensor performs visual scanning, the external axis is rotated to generate visual search points for each side of the box at different angles. The visual search points are calculated by the theoretical welding trajectory model through the transformation relationship and the fixed offset distance of the robot. The set of visual search points for each side is used as the trajectory set.

[0041] The technical solution of the present invention has the following beneficial effects:

[0042] The welding operation of the fuel tank body is quickly constructed by parametric model. When updating the tank body model, only the size information of the tank body needs to be updated to automatically obtain the welding trajectory of the tank body, eliminating the tedious robot teaching work. For visual acquisition, laser vision sensor is used to correct the assembly error and clamping error of the workpiece itself.

[0043] This invention can reduce the difficulty of welding operations, lower the threshold for using robotic arms for welding, and result in high welding efficiency and high weld consistency. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the welding system according to an embodiment of this specification;

[0045] Figure 2 This is a flowchart of the welding method described in the embodiments of this specification;

[0046] Figure 3 A schematic diagram illustrating the external shaft calibration of the welding system as described in this specification.

[0047] Figure 4 This is a schematic diagram illustrating the side length definition of the box body in an embodiment of this specification;

[0048] Figure 5 This is a schematic diagram illustrating the side length definition of the box body in an embodiment of this specification;

[0049] Figure 6 This is a schematic diagram illustrating the side length definition of the box body in an embodiment of this specification;

[0050] Figure 7 This is a schematic diagram illustrating the visual location setting of the enclosure in an embodiment of this specification.

[0051] Figure 8 This is a schematic diagram illustrating the visual location setting of the enclosure in an embodiment of this specification.

[0052] Figure 9 This is a schematic diagram showing the position settings of the box body in an embodiment of this specification;

[0053] Figure 10 This is a schematic diagram showing the location settings of the box in an embodiment of this specification. Detailed Implementation

[0054] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided to make this disclosure more comprehensive and complete, and to fully convey the concept of the example embodiments to those skilled in the art. The described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a full understanding of embodiments of this disclosure. However, those skilled in the art will recognize that the technical solutions of this disclosure can be practiced with one or more of the specific details omitted, or other methods, components, apparatus, steps, etc., can be employed. In other instances, well-known technical solutions are not shown or described in detail to avoid obscuring various aspects of this disclosure.

[0055] like Figure 1 As shown in the figure, this specification provides a teach-free welding system based on a laser vision sensor. The system includes: a robot 101 having multiple degrees of freedom, with a welding torch 102 mounted at its end; an external shaft 103 for fixing a housing 104 and rotating the housing 104 below the robot 101, the fixed housing 104 having a regular shape; a laser vision sensor 105 mounted at the end of the robot 101; and a controller 106 for controlling the operation of the robot 101, the external shaft 103, and the laser vision sensor 105.

[0056] The external shaft 103 can be a positioner, which can be single-axis or multi-axis. The housing can be an oil-immersed transformer tank, or a triangular tank. The laser vision sensor can be mounted on the welding torch.

[0057] Based on the above welding system, such as Figure 2The embodiments of this specification also provide a teach-free welding method based on a laser vision sensor. This welding method is executed by the controller of a welding system and includes steps S201-205:

[0058] In step S201, the normal vector of the robot's external axis relative to the robot's rotation center and rotation plane is calibrated to determine the relationship between the external axis and the robot. The external axis is used to drive the box to rotate in one or more directions, and the box has a regular shape.

[0059] As a supplement, such as Figure 1 and Figure 3 The method involves setting a fixed point mark 108 for calibration on the mounting plate 107 plane of the outer axis for mounting the housing, and driving the outer axis to rotate at multiple angles; obtaining the coordinates of the robot end point pointing to the fixed point mark at different angles, and combining the rotation angle to obtain the rotation center and normal vector of the mounting plate.

[0060] Specifically, the calibration method involves affixing calibrated markers to the surface of the fuel tank mounting plate. By rotating the robot's external axis, three or more positions are taken, i.e., at three or more different J7 angles, with adjacent angle differences between 30° and 40°. The end of the robot tool (welding torch) is then pointed to the calibrated marker on the mounting plate at each J7 angle. Based on the recorded coordinates of the three robot welding torch ends, the rotation center of the mounting plate on the positioner can be calculated. and the normal vector of the rotating plane .

[0061] In step S202, the conversion relationship between the robot end effector and the laser vision sensor is calibrated. The laser vision sensor is used to acquire the vision of the box located on the external axis, and the conversion relationship between the laser vision sensor and the robot end effector is fixed.

[0062] As a supplement, the laser vision sensor is controlled to take pictures of a fixed calibration plate in multiple different postures, and a homogeneous transformation equation is established based on the hand-eye relationship; the transformation relationship is obtained by solving the homogeneous transformation equation.

[0063] Specifically, hand-eye calibration involves calibrating the relationship between the robot tool's end effector and the laser vision sensor. In this invention, the laser vision sensor can be mounted on a welding torch, thus employing a hand-eye calibration method. By controlling the robot end effector to cause the laser vision sensor to photograph a fixed calibration plate in multiple different postures, a system of equations is established based on the hand-eye relationship to obtain the rotation transformation matrix from the laser vision camera coordinate system to the robot end effector coordinate system. This method requires solving the following homogeneous transformation equations:

[0064]

[0065] in, Let A be the rotation transformation matrix from the camera coordinate system to the robot's end effector coordinate system, and B be the rotation transformation matrix for the robot's end effector after two rotations. Based on matrix and linear algebra principles, if there is only one equation, then only the matrix can be solved. If there are two or more equations, then the general solution can be obtained by solving the matrix. The only solution.

[0066] In step S203, a theoretical welding trajectory model of the box is calculated and generated based on the key feature parameters of the box, wherein the key feature parameters include at least the side length of the box.

[0067] As a supplement, the side length of the box is obtained, and when the rotation axis of the external axis coincides with the line connecting the centroid of the top surface and the centroid of the bottom surface of the box, the distance from the centroid of the top surface of the box to each side of its surface is calculated, so as to obtain the position coordinates of each vertex of the box relative to the coordinate system of the robot, and the multiple position coordinates are combined to form the theoretical welding trajectory model.

[0068] Specifically, as an explanation, such as Figure 4 As shown, the key characteristic parameters of the triangular oil tank are defined as follows: long side L1, short side L2, distance between the long and short sides L3, and height L4. Since the rotation axis of the positioner coincides with the line connecting the centroid of the top and bottom planes of the oil tank, the distances from the centroid of the top or top plane of the oil tank to the long and short sides can be calculated, thus obtaining the positions of each vertex of the tank relative to the robot base.

[0069] Considering that the housing can be clamped on an external shaft positioner, and each side can be rotated to the top during welding, it is only necessary to calculate the coordinates of the two endpoints when housing L1 and L2 are at the top.

[0070] According to the enclosure design information, such as Figure 5 As shown, given that angle α is 30°, then , , When the oil tank is installed on the positioner, and L1 is at the top and parallel to the X-axis of the robotic arm, for point p1 on L1 ( ), Using the default robotic arm posture values, the following calculations were obtained:

[0071]

[0072]

[0073]

[0074] For point p2 on L1 ( The calculation yields the following results:

[0075]

[0076]

[0077]

[0078] When the oil tank is installed on the positioner, and L2 is at the top and parallel to the X-axis of the robotic arm, as follows: Figure 6 As shown, for point p3 on L2 ( The calculation yields the following results:

[0079]

[0080]

[0081]

[0082] For point p4 on L2 ( The calculation yields the following results:

[0083]

[0084]

[0085]

[0086] In step S204, trajectory planning is performed based on the theoretical welding trajectory model and the rotation angle of the external shaft to generate a set of trajectories for visual scanning by the laser vision sensor.

[0087] As a supplement, when the laser vision sensor performs visual scanning, the external axis is rotated to generate visual search points for each side of the box at different angles. The visual search points are calculated by the theoretical welding trajectory model through the transformation relationship and the fixed offset distance of the robot, and the set of visual search points for each side is used as the trajectory set.

[0088] Specifically, as explained in this embodiment, the method for visually locating the welding trajectory is as follows: the current straight-line position is calculated by photographing two points on each side of the box, and then the true coordinates of the six endpoints of the box are extracted sequentially by finding the intersection points of adjacent straight lines. Considering that vision can only perform shooting and positioning from directly above the box, it is sufficient to generate the positions of the points on each side of the box that the positioner visually photographs from six different angles.

[0089] First, the positioner rotates to 0 degrees, so that the long side L1 of the workpiece is at the top and relatively horizontal when clamped, generating two visual positioning points for capturing the long side. and Pv1 and Pv2, obtained through step S203, are processed by the hand-eye matrix. The offset distance d along the X-axis of the robotic arm is calculated as follows: Figure 7 As shown.

[0090] The positioner rotates to 60 degrees, at which point the short side L2 of the housing is at the top and relatively horizontal, generating two visual positioning points Pv3 for capturing the short side. ) and Pv4 ( Pv3 and Pv4, obtained through step S203, are processed by the hand-eye matrix. The offset distance d along the X-axis of the robotic arm is calculated as follows: Figure 8 As shown.

[0091] According to the above method, let J be the rotation degree of the positioner, where J can be 0, 60, 120, 180, 240, or 300. From this, the set of points on the six sides of the visual imaging box can be obtained as {Pr1, Pr2, Pr3…Prn} (n=12), as follows:

[0092] Pr1(Pv1,J=0),Pr2(Pv2,J=0),Pr3(Pv3,J=60),Pr4(Pv4,J=60),Pr5(Pv1,J= 120),Pr6(Pv2,J=120),Pr7(Pv3,J=180),Pr8(Pv4,J=180),…Prn(Pv4,J=300)

[0093] In step S205, based on the trajectory set and the transformation relationship, the welding trajectory is calculated when the laser vision sensor scans the weld seam of the box, so that the robot and the external axis can perform linkage welding on the box.

[0094] As an explanation, the set of visual scanning points {Pr1, Pr2…Prn} (n=12) is calculated according to step S204, and then the relationship between the vision and the end of the robotic arm welding torch is considered. The actual coordinate set {Pw1,Pw2…Pwn} (n=12) for visual positioning of the box can be obtained by calculating the target point coordinates for visual recognition.

[0095] To extract the coordinates of the end points of the housing, they can be obtained through the intersection of the two adjacent edges of each end point. For example, when the positioner's rotation degree J=0, as shown... Figure 9As shown, Pw1 and Pw2 form a straight line Line1. By rotating Pr15 and Pr16 60 degrees around the positioner's rotation axis, the current Pw15′ and Pw16′ can be calculated, which form a straight line Line2. Then, the coordinates Pe1 of the intersection point of Line1 and Line2 can be calculated.

[0096] When J=60, as Figure 10 As shown, Pw3 and Pw4 form a straight line Line3. By rotating Pw1 and Pw2 60 degrees around the positioner's rotation axis, we can calculate the current Pw1′ and Pw2′, which form a straight line Line4. Then, we can calculate the coordinates Pe2 of the intersection point of Line3 and Line4.

[0097] When J=120, Pw5 and Pw6 form a straight line Line5. By rotating Pw3 and Pw4 60 degrees around the positioner's rotation axis, we can calculate the current Pw3′ and Pw4′, which form a straight line Line6. Then, we can calculate the coordinates Pe3 of the intersection point of Line5 and Line6.

[0098] Similarly, Pe4, Pe5, and Pe6 can be extracted at J=180, J=240, and J=300, respectively;

[0099] At this point, the set of weld trajectory points {Pe1,Pe2,Pe3,Pe4, Pe5,Pe6} of the tank body relative to the robot base can be calculated. The robot executes the weld trajectory sequence in sequence to complete the welding of the tank body.

[0100] When the specifications of the welded enclosure change, simply repeat steps S203 to S205 to achieve teachless welding of different enclosures.

[0101] As can be seen from the above embodiments, the present invention can quickly construct the welding operation of the tank body through a parametric model. When updating the tank body model, only the size information of the tank body needs to be updated to automatically obtain the welding trajectory of the tank body, eliminating the tedious robot teaching work. In terms of visual acquisition, a laser vision sensor is used, which can correct the assembly error and clamping error of the workpiece itself.

[0102] This invention can reduce the difficulty of welding operations, lower the threshold for using robotic arms for welding, and result in high welding efficiency and high weld consistency.

[0103] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention. In addition, those skilled in the art can understand that although some embodiments described herein include certain features included in other embodiments but not other features, combinations of features of different embodiments are meant to be within the scope of the present invention and form different embodiments. For example, in the above claims, any one of the claimed embodiments can be used in any combination. The information disclosed in this background section is only intended to deepen the understanding of the overall background technology of the present invention, and should not be regarded as an admission or in any form implying that the information constitutes prior art known to those skilled in the art.

Claims

1. A laser vision sensor based teach-less welding method, characterized by, The method comprises: Calibrating a normal vector of a rotation center and a rotation plane of a robot relative to an external shaft of the robot to determine a relationship between the external shaft and the robot, the external shaft being used to rotate a box in one or more directions, the box having a regular shape; during calibration, comprising: setting a calibration point mark on a mounting plate plane of the external shaft for mounting the box, and driving the external shaft to rotate by multiple angles; acquiring coordinates of the robot end pointing to the calibration point mark at different angles respectively, and combining the rotation angles to obtain the rotation center and the normal vector of the mounting plate; Calibrating a conversion relationship between the robot end and a laser vision sensor, the laser vision sensor being used to acquire vision of the box located on the external shaft, and the conversion relationship between the laser vision sensor and the robot end being fixed; Based on key feature parameters of the box, calculating and generating a theoretical welding trajectory model of the box, the key feature parameters at least including side lengths of the box; According to the theoretical welding trajectory model and the rotation angle of the external shaft, performing trajectory planning to generate a trajectory set for the laser vision sensor to perform vision scanning; during trajectory planning, comprising: rotating the external shaft when the laser vision sensor performs vision scanning to generate vision positioning points of each side of the box at different angles, the vision positioning points being calculated by the theoretical welding trajectory model through the conversion relationship and a fixed offset distance of the robot, and taking the vision positioning point set of each side as the trajectory set; According to the trajectory set and the conversion relationship, calculating a welding trajectory when the laser vision sensor scans the box weld to make the robot and the external shaft perform linkage welding on the box.

2. The laser vision sensor based teach-less welding method of claim 1, wherein, The calibration of the conversion relationship between the robot end and the laser vision sensor specifically comprises: Controlling the laser vision sensor to capture a fixed calibration board in multiple different postures, and establishing a homogeneous transformation equation according to a hand-eye relationship; Solving the homogeneous transformation equation to obtain the conversion relationship.

3. The laser vision sensor based teach-less welding method of claim 1, wherein, The calculation and generation of the theoretical welding trajectory model of the box based on the key feature parameters of the box comprises: Acquiring the side lengths of the box, and calculating distances from a top center of gravity of the box to each side of the top face when a rotation axis of the external shaft coincides with a line between a bottom center of gravity and the top center of gravity of the box to obtain position coordinates of each vertex of the box relative to a coordinate system of the robot, and combining multiple position coordinates to form the theoretical welding trajectory model.

4. A teach-less welding system based on a laser vision sensor, characterized in that, Comprises: A robot having multiple degrees of freedom, an end of the robot being mounted with a welding gun; An external shaft being used to fix a box and rotate the box below the robot, the fixed box having a regular shape; A laser vision sensor being mounted on the end of the robot; A controller being used to control work of the robot, the external shaft and the laser vision sensor, and performing the following method during control: The calibration obtains a rotation center of the external shaft relative to the robot and a normal vector of a rotation plane; the calibration comprises: setting a calibration point mark on a mounting plate plane of the external shaft for mounting the box body, and driving the external shaft to rotate by a plurality of angles; obtaining coordinates of the robot end pointing to the calibration point mark at different angles respectively, and combining the rotation angles to obtain the rotation center and the normal vector of the mounting plate; The calibration obtains a conversion relationship between the robot end and the laser vision sensor; Based on key feature parameters of the box body, a theoretical welding trajectory model of the box body is calculated and generated, and the key feature parameters at least include side lengths of the box body; According to the theoretical welding trajectory model and the rotation angle of the external shaft, a trajectory planning is performed to generate a trajectory set for visual scanning of the laser vision sensor; in the trajectory planning, when the laser vision sensor performs visual scanning, the external shaft is rotated to generate visual positioning points of each side of the box body at different angles, the visual positioning points are calculated by the theoretical welding trajectory model through the conversion relationship and a fixed offset distance of the robot, and the visual positioning point set of each side is taken as the trajectory set; According to the trajectory set, the conversion relationship, and when the laser vision sensor scans the box body weld, a welding trajectory is calculated to enable the robot and the external shaft to perform linkage welding on the box body.

5. The laser vision sensor based teach-out-of-sight welding system of claim 4, wherein, The calibration obtains the conversion relationship between the robot end and the laser vision sensor, and specifically comprises: The laser vision sensor is controlled to capture a fixed calibration board in a plurality of different postures, and a homogeneous transformation equation is established according to a hand-eye relationship; The homogeneous transformation equation is solved to obtain the conversion relationship.

6. The laser vision sensor based teach-out welding system of claim 4, wherein, The theoretical welding trajectory model of the box body is calculated and generated based on the key feature parameters of the box body, and comprises: The side lengths of the box body are obtained, and when a rotation axis of the external shaft coincides with a line between a top surface barycenter and a bottom surface barycenter of the box body, distances from the top surface barycenter to each side of the box body are calculated to obtain position coordinates of each vertex of the box body relative to a coordinate system of the robot, and a plurality of position coordinates are combined to form the theoretical welding trajectory model.

Images