Pose Adjustment Of End-effector For Drilling Robot Based On Vision

To improve the precision and efficiency of in-situ precision machining of large parts,this dissertation investigates into a fringe-projection-based measurement system used to assist robot machining.Based the robotic drilling,a measurement scheme of the system is investigated,including registration and localization of robot and workpiece,the normal errors inspection and compensation,and reconstruction of 3D profile.The following contributions have been made:A fringe-projection-based assistant measurement system is proposed,including structure design of measurement system,imaging model and parameter calibration of vision sensor,and 3D reconstruction.In order to measure the pose of robot relative to the workpiece,a method for measuring the localization error of a robot based on binocular vision is proposed.By using the binocular vision of the fringe-projection-based measurement system,the coordinates of the circular reference on the workpiece are obtained.And then the least square method is used to calculate the position and orientation of the robot and workpiece in relation to each other.In order to get the center of the circular reference,a method based on binocular vision is proposed,in which concentric circle and radius of the circle are not needed.To inspect the normal direction at drilled point,a method based on point cloud is studied.Moreover,an alignment method is proposed to compensate the normal error.It is worth pointing out that measuring the normal direction through the 3D reconstruction of the workpiece can replace the special measuring equipment,which improves the flexibility of the measuring system.In this paper,an experimental platform of fringe-projection-based assistant measurement system is built on the basis of TriMule robot.The experimental results show that the accuracy of the assistant measurement system meets the requirements of the robotic drilling,and proposed method is correct and effective.