Study On Calibriation And Control Of Robot Vision Systems For 3D Reconstruction Of Workpiece Surface

When industrial robot vision systems are applied to 3D measurement tasks,for better accuracy,not only the robot vision system needs to be calibrated,but also the camera position needs to be planned and positioned to capture abundant key feature information of a workpiece surface.In the existing measurement method,the camera position needs to be planned by human experience,and the camera positioning is realized by open-loop control of the robot controller.However,this method cannot guarantee the key surface features located in the the camera view,which leads to the lack of key features information and affects the accuracy of 3D reconstruction.Aiming at these problems,this dissertation studies robot vision systems for the 3D reconstruction of workpiece surfaces,and introduces visual servoing into the camera positioning process.Then,the calibration and control mehtod in the system is further developed.The main research contents are listed as follows:For the six-degree-of-freedom industrial robot in the system,a forward kinematic model of the robot is established based on the Denavit–Hartenberg method.Then,the analytical and numerical methods of robot inverse kinematics are developed respectively,and the two methods are verified by simulation experiments.Experimental results guarantee the effectiveness of the methods.For the monocular camera in the system,a camera imaging model is established.Based on Zhang's calibration method,a camera calibration experiment is performed.Then,a classic hand-eye calibration method is introduced.To solve the problems in the method,a hand-eye calibration method based on differential evolution algorithm is proposed.Thereafter,a hand-eye calibration experiment platform is set up and the calibration results of the proposed method and the traditional method are compared.The experimental results verify the effectiveness of the proposed method.A vision servo control framework is established based on the system.To deal with the time delay effect in the system,a time-delay visual servo control method based on the differential quadrature method is proposed.A simulation platform and a physical experiment platform are developed to verify the feasibility and effectiveness of the method.A 3D measurement platform based on a robot vision system is built,and multi-view workpiece images are collected based on the visual servoing method.With the sparse reconstruction method and the dense reconstruction method,the 3D reconstruction experiment of the workpiece surface is performed.The experimental results verify the effectiveness of the robot vision system calibration and control method for 3D reconstruction of workpiece surfaces proposed by this dissertation.