Research On Motion Parameter Identification Of Manipulator Based On Computer Vision

With the rapid development of computer vision technology,stereo vision measurement technology has also been widely used in aerospace engineering practice.Especially in the large space manipulator system,the use of visual measurement technology to identify the position,attitude and motion parameters of the robot arm have become the focus of research,especially in the large space.By in-depth studying in visual measurement technology in this paper,I designed a 3D measurement system.It is implemented to measure a set of motion parameters of mechanical arm system,and this system can effectively identify the gear space manipulator system.First,the theory of binocular vision calibration technology is studied in detail.At the same time,I calculate the distortion parameters of the camera internal parameters and the relative pose between two cameras,using the Zhang Zhengyou plane.In order to get the motion parameters of the manipulator,I need to track the target point.The traditional Mean-Shift algorithm was improved by object tracking.By tracking the target area,extracting the edge of the image,ellipse fitting of the edge points and extracting the center points,the accurate tracking of the target image position is achieved.Finally,using the mathematical model of 3D measurement,the space point coordinates of the tracked image feature points are calculated,and the kinematic parameters of the robot arm are measured.The backlash affects the movement of the robot arm and the service life of the robot.By researching the driving wheel in low speed and small angle oscillation,the stagnation phenomenon of driven wheel after separation from driving wheel,the backlash of mechanical arm system is calculated after measuring the dead time and rotation velocity.In this paper,the positions of the point at end and mid of the mechanical arm are measured respectively,and the rotation angle of the driving wheel and driven wheel is measured respectively.The simulation results of the experimental data and the 3D simulation model are compared.The similarity coefficient of Pearson is more than 95%.The feasibility of the 3D model simulation is verified.The relative error between the measured tooth clearance and the theoretical value obtained by experimental measurement is about 5%.With in the range of allowable error,the data is proved by the calculation theory.The experimental results show that the binocular vision measurement method has high practical value for the motion state of the robot arm and the measurement of the gear clearance.