Robot welding gun tool automatic calibration device deviation correction method of binocular vision sensor

By using image recognition and processing algorithms from a binocular vision sensor, automatic calibration and additional axis error correction of the welding torch tool are achieved, solving the problems of high cost and low efficiency. It is applicable to various types of welding torches and improves welding accuracy and efficiency.

CN115213611BActive Publication Date: 2025-11-18安徽工布智造工业科技有限公司
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202211065061.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-08-31
Publication Date
2025-11-18
Estimated Expiration
2042-08-31

AI Technical Summary

Technical Problem

In the existing technology, the automatic calibration of welding torch tools requires the assistance of high-cost precision instruments and cannot correct the return error of the additional shaft, resulting in low welding efficiency.

Method used

The automatic calibration device for robotic welding torches, which uses binocular vision sensors, automatically corrects deviations by recognizing the position of the welding torch sleeve and welding wire through image recognition and combining image processing algorithms to calculate offset values ​​and additional axis errors.

Benefits of technology

It reduces calibration costs, is applicable to various welding torch types, reduces manpower and material consumption, improves welding accuracy and efficiency, and avoids welding misalignment problems during long welding processes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115213611B_ABST
    Figure CN115213611B_ABST
Patent Text Reader

Abstract

The present application relates to the field of machine vision, especially to a robot welding gun tool automatic calibration device deviation correction method of binocular vision sensor, the robot welding gun tool automatic calibration device of binocular vision sensor comprises: a workbench, an upper monocular camera, a side monocular camera and a welding gun, the present application only adopts two industrial cameras, and calibrates by using an image processing algorithm, which has the characteristics of low cost and not easy to damage, the present application aims at the problem that the tool coordinate system is inaccurate due to deformation of the welding gun tool in the working process under long-time high temperature and other environments, obtains the offset value of the welding gun tool in the image coordinate system by using the image processing mode based on the binocular vision sensor, thereby obtaining the offset value in the actual space, and then modifying the tool coordinate system, thereby avoiding the problem of offset welding in the long-time welding process.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of machine vision, and more particularly to a method for correcting deviations in an automatic calibration device for a robotic welding torch tool using a binocular vision sensor. Background Technology

[0002] With the rapid development of the intelligent welding field, intelligent welding robots have also been widely used. Currently, welding torches are often calibrated and corrected manually, but this is time-consuming and affects welding efficiency. Intelligentization and automation are inevitable development directions for the manufacturing industry; therefore, automatic tool calibration is an inevitable trend. Based on this, adopting automatic tool calibration is an effective solution to save manpower and resources.

[0003] Existing automatic tool calibration technology typically involves rotating the robot's end effector within a precision instrument to obtain the precise spatial positioning of the end effector and the current robot posture, thereby obtaining the accurate TCP value. However, automatic tool calibration based on precision instruments often requires high-cost precision instruments for assistance and cannot correct for additional axis backlash. Summary of the Invention

[0004] Therefore, this invention was made in view of the above problems. The purpose of this invention is to solve the problems of high cost and inability to correct the return error of the additional axis by using an automatic calibration device for robotic welding gun tools with binocular vision sensors. This invention achieves the above objective through the following technical solution:

[0005] 1. A method for correcting deviations in an automatic calibration device for a robotic welding torch tool using a binocular vision sensor, characterized in that: the automatic calibration device for a robotic welding torch tool using a binocular vision sensor includes: a work platform, an upper monocular camera, a side monocular camera, and a welding torch;

[0006] The correction method is as follows:

[0007] S1: First, locate the welding torch sleeve position; obtain the position of point B on the welding torch sleeve through image recognition, and point O is the center point of the upper monocular camera. Draw the foot of the perpendicular from point B on the welding torch sleeve to OH to obtain the line BH. The angle α between OB and BH can then be calculated:

[0008]

[0009]

[0010] S2: Position the welding wire; C is the center of the welding gun, i.e. the contact tip point, F is the welding wire point, the welding wire length CF, CD is the length of the welding wire blocked by the welding gun in the field of view of the upper monocular camera, the angle between the welding gun and the working platform (1) is β, and A is the point on the plane from the welding wire CF to the welding gun sleeve point B in the field of view of the camera;

[0011] Given CD = tan(α - (90 - β)) * BC, then BD can be calculated:

[0012]

[0013] Draw a line parallel to CF through point B, intersecting AF at point G. Then we can find BG: OB / (BD+OB)=BG / (CF-CD)

[0014] Draw a line through G parallel to BD, intersecting CF and E, such that ∠GEF = ∠BDF = α + β. Then DE = BG and EF = CF - CD - DE. According to the Law of Cosines, we can find:

[0015]

[0016] Then ∠FGE can be calculated:

[0017] FE / sin(∠FGE)=GF / sin(∠GEF)

[0018] That is, ∠AOB = ∠FGE

[0019] The length of AB can then be calculated:

[0020] AB / sin(∠AOB)=OB / sin(∠OAB) ∠OAB=180-∠AOB-∠α

[0021] Therefore, based on the distance AB from the welding wire projection point A to the welding gun sleeve point B, the position of the welding wire CF under the camera can be obtained.

[0022] Coordinate values;

[0023] S3: Calculate the offset value; according to the user-defined program, the welding torch moves in a circle, passing through several points. A circle is obtained by fitting these points together. In the diagram, point K is the reference standard value, point M is the point obtained after the welding torch offset, MN is the straight line containing the welding wire, and N is the foot of the perpendicular from the standard point K to the straight line MN. Since the straight line containing MN and point K are known, point N can be calculated. Therefore, the lengths of KN and MN can be obtained. The lengths of KN and MN are the offset values ​​in the X and Y directions of the welding torch offset.

[0024] By combining calculations at several points on the circumference, the error value can be reduced;

[0025] S4: Correct for additional axis error; When only welding torch offset occurs, the center of the circle will not shift. However, when the center shifts, it must be due to the backlash difference of the additional axis. As shown in Figure 5, the welding torch 4 moves around a circle, passing through several points. A circle is obtained by fitting these points. Points r1 and r2 are the center points. The cross center lines of the two circles are drawn through r1 and r2 respectively. The horizontal center line through r1 and the vertical center line through r2 intersect at point L. r1L and r2L are the offset distances in the longitudinal and transverse directions, respectively. Since points r1 and r2 are known, the distances r1L and r2L can be calculated, which are the backlash differences caused by the interpolation additional axis in the Y and X directions, respectively.

[0026] Preferably, the upper monocular camera is located at the upper end of the working platform 1 and is fixed on the working platform.

[0027] Preferably, the side monocular camera is located on the side of the working platform and is also fixed on the working platform (1).

[0028] Preferably, the welding torch is located below the upper monocular camera, the upper monocular camera (2) detects the changes in the x and y directions (left and right) of the welding torch, and the side monocular camera detects the changes in the z direction (height) of the welding torch.

[0029] Beneficial effects of this invention:

[0030] 1. The tool calibration method of the present invention using a binocular vision sensor can be adapted to all types of welding torches, and is not limited to a single tool;

[0031] 2. This invention uses only two industrial cameras and employs image processing algorithms for calibration. It is characterized by low cost and low damage resistance, thereby avoiding the manpower and material resources consumed by frequent manual calibration of welding torches in industrial settings, making it possible for some small and medium-sized enterprises to introduce this device.

[0032] 3. This invention addresses the problem of inaccurate tool coordinates caused by deformation of welding torches during prolonged operation in high-temperature environments. Based on a binocular vision sensor, image processing is used to obtain the offset value of the welding torch in the image coordinate system, thereby obtaining the offset value in actual space. The tool coordinate system is then modified to avoid the problem of misaligned welding during long-term welding. Attached Figure Description

[0033] Figure 1 is a schematic diagram of the overall structure of the present invention.

[0034] Figure 2 is a front view of the welding torch and the upper monocular camera of the present invention.

[0035] Figure 3 is a partially enlarged view of the welding torch and the upper monocular camera of the present invention.

[0036] Figure 4 is a schematic diagram of the calculation of the offset value according to the present invention.

[0037] Figure 5 is a schematic diagram of the additional shaft error correction of the present invention. Attached Figure Description

[0039] 1. Working platform; 2. Top monocular camera; 3. Side monocular camera; 4. Welding torch. Detailed Implementation

[0040] Preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, the present invention can also be implemented in various different forms, and therefore the present invention is not limited to the embodiments described below. In addition, for the purpose of more clearly describing the present invention, parts not connected to the invention will be omitted from the drawings.

[0041] This invention uses a binocular vision sensor for calibration to correct welding torch offset errors on the welding site and add...

[0042] Axis error correction enables the binocular vision sensor to more accurately detect the welding position and perform welding.

[0043] As shown in Figure 1, the automatic calibration device for robot welding torch tools with binocular vision sensor includes an automatic calibration device for robot welding torch tools with binocular vision sensor. The automatic calibration device for robot welding torch tools with binocular vision sensor includes: a working platform 1, an upper monocular camera 2, a side monocular camera 3, and a welding torch 4.

[0044] The working platform 1 serves to fix and support other components;

[0045] The upper monocular camera 2 is located at the upper end of the working platform 1 and is fixed on the working platform 1;

[0046] The side monocular camera 3 is located on the side of the working platform 1, and the side monocular camera 3 is also fixed on the working platform 1;

[0047] The welding torch 4 is located below the upper monocular camera 2. The upper monocular camera 2 detects the changes in the x and y directions (left and right) of the welding torch 1, and the side monocular camera 3 detects the changes in the z direction (height) of the welding torch 1.

[0048] This invention also discloses a method for correcting deviations in an automatic calibration device for robotic welding torches using binocular vision sensors, the steps of which are as follows:

[0049] As shown in Figures 2, 3, and 4, the error correction calculation for the welding torch offset value is performed.

[0050] S1: As shown in Figure 2, first position the welding torch sleeve;

[0051] The position of point B on the welding torch sleeve is obtained through image recognition. Point O is the center point of the upper monocular camera 2. The foot of the perpendicular is drawn from point B on the welding torch sleeve to OH, and the angle α between line BH and line OB and BH can be obtained. Then, OB can be calculated as: tan(α) = OH / BH

[0052]

[0053] S2: As shown in Figures 2 and 3, position the welding wire.

[0054] Point C is the center of welding torch 4, i.e., the contact tip position; F is the welding wire point; the welding wire length is CF; CD is the length of welding wire obscured by welding torch 4 within the field of view of the upper monocular camera 2; the angle between welding torch 4 and working platform 1 is...

[0055] β, point A is a point on the plane from welding wire CF to welding gun sleeve point B within the camera's field of view;

[0056] Given CD = tan(α - (90 - β)) * BC, then BD can be calculated:

[0057]

[0058] Draw a line parallel to CF through point B, intersecting AF and G. Then BG can be found.

[0059] OB / (BD+OB)=BG / (CF-CD) and ∠GEF=∠BDF= α+β

[0060] Draw a line through G parallel to BD, intersecting CF and E. Then DE = BG and EF = CF - CD - DE.

[0061] Then, according to the Law of Cosines, we can find that:

[0062]

[0063] Then ∠FGE can be calculated:

[0064] FE / sin(∠FGE)=GF / sin(∠GEF)

[0065] That is, ∠AOB = ∠FGE

[0066] The length of AB can then be calculated:

[0067] AB / sin(∠AOB)=OB / sin(∠OAB) ∠OAB=180-∠AOB-∠α

[0068] Therefore, the coordinates of the welding wire CF under the camera can be obtained from the distance AB from the projection point A of the welding wire to the point B of the welding gun sleeve.

[0069] S3: As shown in Figure 4, calculate the offset value;

[0070] According to the pre-defined program, the welding torch 4 moves in a circle, passing through several points. A circle is obtained by fitting these points together. In the diagram, point K is the reference standard value, point M is the point obtained after the welding torch 4 is offset, MN is the straight line containing the welding wire, and N is the foot of the perpendicular from the reference point K to the straight line MN. Since the straight line containing MN and point K are known, point N can be calculated, and thus the lengths of KN and MN can be obtained. The lengths of KN and MN are then the offset values ​​of the welding torch 4 in the X and Y directions. Therefore, the circle...

[0071] By performing comprehensive calculations on several points, the error value can be reduced;

[0072] S4: As shown in Figure 5, perform correction of additional shaft error;

[0073] When only the welding torch deflects, the center of the circle does not shift. However, when the center shifts, it is necessarily due to the backlash difference of the additional axis. As shown in Figure 5, the welding torch 4 moves around a circle, passing through several points. A circle is obtained by fitting these points. Points r1 and r2 are the center points. The cross center lines of the two circles are drawn through r1 and r2 respectively. The horizontal center line through r1 and the vertical center line through r2 intersect at point L. r1L and r2L are the longitudinal and lateral offset distances, respectively. Since points r1 and r2 are known, the distances r1L and r2L can be calculated, which are the backlash differences caused by the interpolation additional axis in the Y and X directions, respectively.

Claims

1. A method for correcting deviations in an automatic calibration device for a robotic welding torch tool using a binocular vision sensor, characterized in that: The automatic calibration device for the robotic welding torch tool of the binocular vision sensor includes: a working platform (1), an upper monocular camera (2), a side monocular camera (3), and a welding torch (4). The correction method is as follows: S1: First, locate the sleeve of the welding torch (4); obtain the position of point B of the welding torch sleeve through image recognition, point O is the center point of the upper monocular camera (2), OH is a straight line drawn vertically downward from the origin of the camera view coordinate system O, and intersects the upper edge B of the welding torch sleeve protective sleeve at point H on the same horizontal plane. Draw the foot of the perpendicular from point B of the welding torch sleeve to OH to obtain the straight line BH. The angle α between OB and BH can be calculated: ; S2: Position the welding wire; Point C is the center of the welding gun (4), i.e. the contact tip point; F is the welding wire point; the welding wire length is CF; Point D is obtained by connecting the origin O of the camera view coordinate system with the upper edge B of the welding gun sleeve protective sleeve and intersecting the extension line CF of the sleeve at point D; CD is the length of the welding wire blocked by the welding gun (4) in the camera field of view of the upper monocular camera (2); the angle between the welding gun (4) and the working platform (1) is β; Point A is the point on the plane from the welding wire CF to the welding gun sleeve point B in the camera field of view; Given CD = tan(α - (90 - β)) * BC, then BD can be calculated: ; Draw a line parallel to CF through point B, intersecting AF at G. Then we can find BG: OB / (BD+OB)=BG / (CF-CD). Draw a line parallel to BD through G, intersecting CF at E, and ∠GEF=∠BDF=α+β. Then DE=BG, EF=CF-CD-DE. Then, according to the Law of Cosines, we can find that: ; Then ∠FGE can be calculated: FE / sin(∠FGE)=GF / sin(∠GEF) That is, ∠AOB = ∠FGE The length of AB can then be calculated: AB / sin(∠AOB) = OB / sin(∠OAB) ∠OAB = 180 - ∠AOB - ∠α Therefore, the coordinates of the welding wire CF under the camera can be obtained based on the distance AB from the welding wire projection point A to the welding gun sleeve point B; S3: Calculate the offset value; according to the program you set, the welding gun (4) moves around a circle and passes through several points. The circle is obtained by fitting the several points. Point K is the standard point, point M is the point obtained after the welding gun (4) is offset, MN is the straight line where the welding wire is located, and N is the foot of the perpendicular from the standard point K to the straight line MN. Since the straight line where MN is located is known and point K is known, point N can be calculated. Then the lengths of KN and MN can be obtained. The lengths of KN and MN are the offset values ​​of the welding gun (4) in the X and Y directions. The error value is reduced by comprehensively calculating the several points on the circle. S4: Correct the error of the additional axis; when only the welding gun offset occurs, the center of the circle will not be offset. However, when the center of the circle is offset, it must be due to the backlash difference of the additional axis. The welding gun (4) moves around the circle and passes through several points. A circle is obtained by fitting these points. Points r1 and r2 are the center points. The cross center lines of the two circles are drawn through r1 and r2 respectively. The horizontal center line through r1 and the vertical center line through r2 intersect at point L. r1L and r2L are the offset distances in the longitudinal and transverse directions, respectively. Since points r1 and r2 are known, the distances of r1L and r2L can be calculated, which are the backlash caused by the interpolation additional axis in the Y and X directions, respectively.

2. The automatic calibration method for robotic welding torch tools using a binocular vision sensor according to claim 1, characterized in that: The upper monocular camera (2) is located on the upper end of the working platform (1) and is fixed on the working platform (1).

3. The automatic calibration method for robotic welding torch tools using a binocular vision sensor according to claim 1, characterized in that: The side monocular camera (3) is located on the side of the working platform (1) and is also fixed on the working platform (1).

4. The automatic calibration method for robotic welding torch tools using a binocular vision sensor according to claim 1, characterized in that: The welding torch (4) is located below the upper monocular camera (2). The upper monocular camera (2) detects the changes in the x and y directions of the welding torch (4), and the side monocular camera (3) detects the changes in the z direction of the welding torch (4).

Citation Information

Patent Citations

  • Device and method for calibrating welding gun tool point of welding robot and workpiece coordinate system

    CN108393896A

  • Method for automatically measuring and correcting welding gun gesture of welding robot

    CN111299929A