A device for evaluating the alignment accuracy of a mechanical arm hole axis based on binocular vision guidance

Abstract

The application discloses a device for evaluating the hole axis alignment precision of a mechanical arm based on binocular vision guidance. For the evaluation of the hole axis alignment precision of the mechanical arm guided by binocular vision, the application designs an evaluation device, which is divided into two parts. One part is a calibration board with a circular chessboard pattern for evaluating the positioning precision of binocular vision. The position and posture of the circular mark point are determined by establishing the calibration board coordinate system, which serves as the evaluation true value of this part. The other part is a standard device for evaluating the overall positioning precision of the vision-guided mechanical arm system. The standard device is composed of three holes with cone tip structures, wherein the cone tip and the upper end surface of the hole are concentric. The position coordinates of the hole in the base coordinate system of the mechanical arm are obtained by aligning the cone tip at the end of the mechanical arm with the cone tip on the standard device. The posture is obtained by the three concentric cone tips, which serves as the evaluation true value of this part. The designed device is used for evaluating the binocular vision algorithm and the hole axis alignment precision of the binocular vision-guided mechanical arm.

Description

Technical Field

[0001] This invention belongs to the field of automated assembly of vision-guided robotic arms, and in particular relates to a device for evaluating the hole-axis alignment accuracy of binocular vision-guided robotic arms. Background Technology

[0002] In recent years, the research and application of robotic arm assembly have been widely developed in various fields, such as the assembly of large aerospace components and aircraft parts. The precision consistency, automation level, and effectiveness of robotic arm assembly have been verified in many complex industrial environments. Hole-axis alignment is the most common form of robotic arm assembly. Mating components include shafts and holes. To improve the flexibility and accuracy of robotic arm shaft-hole alignment systems, vision sensors, force sensors, and hybrid sensors are increasingly being integrated into these systems to provide feedback information for robotic arm posture adjustment and coaxial hole-axis alignment. Compared to other methods, vision sensors are non-contact and information-rich, and can be integrated into robot end effectors to obtain the target workpiece's posture, physical dimensions, and surface reconstruction.

[0003] Because the alignment accuracy requirements for holes and shafts in industrial applications are generally high, the accuracy of vision-guided robotic arms in automating hole and shaft alignment is difficult to quantify. Direct assembly inevitably damages the workpiece. A common solution is to use force sensors to provide feedback on the force on the shaft to determine whether it is aligned. However, this contact-based assessment may damage the workpiece or even the robotic arm. Alternatively, a laser tracker can be used to calibrate the hole, shaft, and robotic arm from all angles to quantify the alignment accuracy. However, this non-contact assessment method based on laser trackers is complex to operate and expensive.

[0004] Therefore, to solve the above problems, this invention aims to design an evaluation device that is simple to operate and easy to install, and can complete the evaluation of positioning accuracy solely relying on the system itself. A novel evaluation method and device design are proposed. Utilizing an innovative mechanical structure design and a corresponding simple and effective evaluation method, the device achieves the evaluation of the hole-axis alignment accuracy of a binocular vision-guided robotic arm. This provides data support for subsequent error compensation, which is beneficial for improving the positioning accuracy and reliability of automated assembly of vision-guided robotic arms, and provides a new solution for the accuracy evaluation of hole-axis alignment, a common and precise assembly operation. Summary of the Invention

[0005] This invention relates to an innovative device for evaluating the hole-axis alignment accuracy of a binocular vision-guided robotic arm. It aims to assess the reliability of the vision-guided robotic arm during automated assembly and can also serve as a compensation input to improve assembly accuracy. The device consists of two parts: a spatial circle positioning accuracy evaluation device to assess the accuracy and reliability of the binocular vision spatial circle pose estimation algorithm, and an overall hole-axis alignment accuracy evaluation device for the binocular vision robotic arm system. Furthermore, this device can achieve high-precision calibration of the robotic arm's end effector. The following is a detailed description of the invention:

[0006] Unlike traditional calibration boards, this calibration board features circular markers, is made of optical glass, and has an adjustable backlight. This innovative design allows for the provision of high-quality calibration board images in various working environments. Furthermore, the three-dimensional pose of the circular markers can be determined using the calibration board coordinate system, which is used to evaluate the accuracy and reliability of the binocular vision spatial circular positioning algorithm.

[0007] The standard consists of three hole structures with tapered tips. This innovative design can not only calibrate the end-effector of a robotic arm using a single tapered tip, but also use the three tapered tips as spatial position coordinates to determine the true value of the spatial pose of the hole, so as to evaluate the accuracy and reliability of automated hole alignment.

[0008] Images of the calibration board in different poses are acquired. Camera calibration is used to obtain the pose relationship between the calibration board and the camera, thereby establishing a transformation matrix between the calibration board coordinate system and the camera coordinate system. The calibration board coordinate system is established at the upper left corner. The position and pose of the circular marker point can be determined in the calibration board coordinate system by design parameters. The circular marker point images acquired by the left and right cameras are identified by the binocular vision spatial circular positioning algorithm to obtain its pose in the camera coordinate system. Using the transformation matrix between the calibration board coordinate system and the camera coordinate system established above, the measured pose of the circular marker point is transformed to the calibration board coordinate system and compared with the true value obtained by calibration to obtain the binocular vision spatial circular pose positioning accuracy.

[0009] By manually or automatically aligning the end effector cone tip calibration tool of the robotic arm with the cone tip of the standard, the three-dimensional position coordinates of the cone tip inside the hole in the robotic arm's base coordinate system are obtained. By measuring the position coordinates of the three cone tips on the standard, the normal of the plane containing the hole end face can be determined, thereby determining the position and orientation of the hole in the robotic arm's base coordinate system. Through hand-eye calibration and TCP calibration, the transformation relationship matrix from the camera coordinate system to the end effector coordinate system of the robotic arm and then to the robotic arm's base coordinate system can be obtained. By using the robotic arm to drive the binocular camera to identify and locate the hole on the standard, its pose in the camera coordinate system is obtained. Using the above transformation matrix, it is transformed to the robotic arm's base coordinate system and compared with the true value obtained from calibration, thereby obtaining the hole axis alignment accuracy of the binocular vision-guided robotic arm.

[0010] When performing TCP calibration and determining the true value of hole pose, an auxiliary tool was designed at the end of the robotic arm. This auxiliary tool is a conical structure that can be fixed by tightening or gluing, which facilitates quick installation and disassembly. The conical structure also facilitates operation by determining the calibration position.

[0011] The calibration tool is designed with user convenience in mind, making it suitable for experienced operators as well as beginners who can quickly learn and master it.

[0012] Compared with the prior art, the present invention is characterized by the fact that it can achieve the evaluation of positioning accuracy by using only the system itself without relying on other third-party sensors, and the evaluation device can be adjusted according to the actual working conditions, including adjusting the brightness of the calibration plate and the specifications and combination of the cone tip on the standard to adapt to the needs of different scenarios.

[0013] In summary, this invention provides a holistic solution that, by combining simplified hardware design with intelligent software processing, aims to improve the reliability of automated assembly using binocular vision-guided robotic arms, and further provides calibration data to improve the accuracy of automated assembly. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the principle of a device for evaluating the alignment accuracy of a hole shaft in a binocular vision-guided robotic arm, as implemented in this invention.

[0015] Part numbers in the diagram: 1-robotic arm, 2-calibration plate with circular checkerboard, 3-standard device with conical tip, 4-auxiliary conical tip at the end of the robotic arm.

[0016] Figure 2 For calibration plates with a circular checkerboard pattern.

[0017] Part number in the picture: 2 - Calibration plate with circular checkerboard pattern.

[0018] Figure 3 It is a standard with a conical tip.

[0019] Part number in the diagram: 3 - Standard with cone tip.

[0020] Figure 4 It serves as an auxiliary cone tip at the end of the robotic arm.

[0021] Part number in the picture: 4-Assisted cone tip at the end of the robotic arm.

[0022] Figure 5 This is a flowchart of a device used to evaluate the hole alignment accuracy of a binocular vision-guided robotic arm. Detailed Implementation

[0023] In order to achieve the aforementioned invention and solve the problems existing in the background art, the present invention proposes the following specific embodiments:

[0024] 1. Design and manufacture of calibration plates with circular checkerboard patterns:

[0025] The design uses a checkerboard calibration plate (2) with a backlight and made of optical glass. The illuminance of the light source can be continuously adjusted within the range of 0-3000lx, and the calibration plate has a processing accuracy of ±0.001mm.

[0026] 2. Design and manufacture of standards with tapered tips:

[0027] A standard with a tapered tip (3) is manufactured using high-precision machining technology to ensure that the shape, angle and size of the tip meet the design specifications, thereby providing an accurate calibration reference.

[0028] The selection of high-quality materials and the heat treatment process for advanced tools ensure the stability of the tip shape under different temperatures and usage conditions.

[0029] 3. Design and manufacture of auxiliary cone tips for the end effector of robotic arms:

[0030] The auxiliary cone tip structure at the end of the robotic arm is manufactured using high-precision machining technology (4) to ensure that the shape, angle and size of the tip meet the design specifications, thereby providing an accurate calibration reference.

[0031] The selection of high-quality materials and the heat treatment process for advanced tools ensure the stability of the tip shape under different temperatures and usage conditions.

[0032] 4. System Calibration:

[0033] Twenty images of the designed checkerboard calibration board (2) were collected to calibrate the internal and external parameters and distortion parameters of the left and right cameras, providing a benchmark for subsequent image calibration. Based on the six-point calibration principle, the designed standard with a cone tip (3) and the auxiliary cone tip (4) at the end of the robotic arm were used to complete the TCP calibration of the robotic arm. The robotic arm was operated to drive the binocular camera fixed at the end joint of the robotic arm to collect images of the calibration board and record the TCP coordinates at the end. The Tsai calibration method was used to complete the hand-eye calibration. Thus, the transformation relationship from the camera coordinate system to the end tool coordinate system of the robotic arm and then to the base coordinate system of the robotic arm was determined.

[0034] 5. Evaluation of the accuracy of binocular vision spatial circular positioning:

[0035] The circular feature points on the 20 images collected for calibration were identified, and the calibration results were used to transform them into the calibration board coordinate system. The circular pose was compared with that in the actual designed calibration board coordinate system to obtain the binocular vision spatial circle positioning accuracy. For the reliability of the results, the average error of the 20 calculation results was taken as the evaluation result.

[0036] 6. Determination of the true pose of the hole in the robot arm's base coordinate system:

[0037] The standard with a cone tip (3) is fixed in place. The robotic arm (1) is manually operated to drive the end auxiliary cone tip (4) to coincide with the cone tip (3) of the standard. The three-dimensional coordinates of the cone tip in the base coordinate system of the robotic arm are read from the teach pendant. This operation is performed on all three cone tips on the standard. Each cone tip is operated 10 times. The average value is taken as the true value of the hole position. The normal of the plane where the cone tip and the hole are located can be determined through the three cone tips. Thus, the true value of the hole position and orientation is determined.

[0038] 7. Evaluation of the alignment accuracy of the hole axis in a binocular vision-guided robotic arm:

[0039] The standard with a cone tip (3) is fixed in place. The robotic arm (1) is manually operated to drive the binocular camera to collect, identify and measure the hole with a cone tip inside the standard. The pose coordinates of the hole in the camera coordinate system are obtained. The transformation relationship matrix from the camera coordinate system to the robotic arm base coordinate system obtained by the above calibration is used to unify the measured pose to the robotic arm base coordinate system. The hole axis alignment accuracy is obtained by comparing it with the true pose value obtained above. In order to ensure the reliability of the result, each hole is measured ten times and the average value is taken as the measurement result.

[0040] Through the above embodiments, this invention not only provides an evaluation device and method for the automated hole and shaft alignment accuracy of binocular vision-guided robotic arms, but also provides calibration data for further improving alignment accuracy. Furthermore, the device is portable, easy to operate, and highly adaptable to various environments. This invention is of great significance for the field of high-precision automated assembly using vision-guided robotic arms.

Claims

1. A device for evaluating the hole-axis alignment accuracy of a binocular vision-guided robotic arm, characterized in that, The system comprises the following components: a robotic arm (1), a calibration plate with a circular checkerboard pattern (2), a standard with a conical tip (3), an auxiliary conical tip at the end of the robotic arm (4), and a binocular camera (5). The robotic arm (1) carries the binocular camera (5) and the auxiliary conical tip at the end of the robotic arm (4) and is the actuator of the entire system. The calibration plate with a circular checkerboard pattern (2) is placed on a marble base to calibrate the binocular camera (5) and evaluate its positioning accuracy. The standard with a conical tip (3) is a standard component with a coaxial cone-hole composite structure, and its interior is equipped with a coaxial... A cone and a cylindrical hole; the cone tip is located at the center point of the upper end face of the cylindrical hole, and the bottom surface of the cone and the bottom surface of the cylindrical hole are on the same plane; this structure makes the cone tip of the cone serve as both the center reference point of the hole end face and the matching reference for positioning of the auxiliary cone tip (4) at the end of the robotic arm; the auxiliary cone tip (4) at the end of the robotic arm is fixed at the end joint of the robotic arm (1) and used in conjunction with the cone tip standard (3); a binocular camera (5) is mounted at the end joint of the robotic arm (1) for system identification, positioning and guidance.

2. The apparatus for evaluating the hole-axis alignment accuracy of a binocular vision-guided robotic arm according to claim 1, characterized in that, The calibration plate with a circular checkerboard pattern (2) differs from the traditional calibration plate in that it has a circular marker on the right side, is made of optical glass, and has a uniform surface light source as its backlight source. At the same time, the illuminance of the light source is adjustable from 0 to 3000 lx, which can adapt to different measurement environments.

3. The apparatus for evaluating the hole-axis alignment accuracy of a binocular vision-guided robotic arm according to claim 1, characterized in that, The cone tip on the standard with cone tip (3) can be replaced with different size combinations to evaluate holes of different specifications.

4. The apparatus for evaluating the hole-axis alignment accuracy of a binocular vision-guided robotic arm according to claim 1, characterized in that, The method of using this device includes the following steps: The binocular visual stereo calibration is achieved using the circular checkerboard calibration plate (2), the camera intrinsic and extrinsic parameters, distortion parameters and the transformation relationship of the left and right camera coordinate systems are determined, and the checkerboard coordinate system is established to obtain the three-dimensional pose of the circle as the true value. The standard with a cone tip (3) and the auxiliary cone tip (4) at the end of the robotic arm are used to complete the calibration of the end tool of the robotic arm and establish the transformation relationship from the coordinate system of the end tool of the robotic arm to the base coordinate system of the robotic arm. Fix the calibration plate (2) with a circular checkerboard pattern, operate the robotic arm (1) to drive the binocular camera (5) to collect images of the calibration plate in different postures, construct the hand-eye relationship matrix, and solve the transformation relationship from the camera coordinate system to the end tool coordinate system of the robotic arm; The binocular camera (5) identifies and locates the circle on the circular checkerboard calibration plate (2) to obtain the measured value. By comparing it with the true value, the binocular vision spatial circle positioning accuracy is obtained. The image acquired during calibration is identified, and the average error is taken as the evaluation result. Place the standard with a cone tip (3) in a suitable position within the working range of the robotic arm (1). Manually operate the robotic arm to align the auxiliary cone tip (4) at the end of the robotic arm with the cone tip of the standard with a cone tip (3). Record the coordinates of the auxiliary cone tip at the end of the robotic arm in the base coordinate system at this time. Repeat the operation of aligning the auxiliary cone tip (4) at the end of the robotic arm with the cone tip of the standard with a cone tip (3) ten times for each of the three cone tips on the standard. Record the coordinates and take the average value as the measurement result. Finally, calculate the normal of the plane in which the three cone tips are located by the three-dimensional position coordinates of the three cone tips. Thus, the position and orientation of the cone tips are obtained as the true value of the hole orientation. Manually operate the robotic arm (1) to drive the binocular camera (5) so that the standard with the cone tip (3) is within the field of view of the binocular camera (5). The binocular camera (5) is triggered by the synchronous trigger to synchronously collect and identify the holes on the standard, and the measured value of the hole pose is obtained. Repeat the operation of the robotic arm to drive the binocular camera (5) to collect the standard image in different poses and obtain the measured value of the hole pose. Compare it with the true value of the hole pose and take the average error as the evaluation result.

5. The apparatus for evaluating the hole-axis alignment accuracy of a binocular vision-guided robotic arm according to claim 1, characterized in that, It includes a set of binocular vision spatial circle localization algorithms, which can solve the spatial circle pose by acquiring target images through binocular cameras (5).

6. The apparatus for evaluating the hole-axis alignment accuracy of a binocular vision-guided robotic arm according to claim 5, characterized in that, The binocular vision spatial circle positioning algorithm is integrated into a host computer, which also has functions for camera parameter setting, image acquisition and storage, and recognition result output.

Images