3D Positioning and Robot Assembly Method for Multi-plane Workpiece

Aiming at the assembly of workpieces with multiple plane structures, this paper proposes a 3D visual positioning and robot assembly method. First, several planes are extracted by plane detection and outlier removal algorithms; then the poses of the plane point clouds are calculated by the oriented bounding box algorithm; finally, the target plane is found through the length and width constraints. Based on proper hand-eye calibration and motion planning, the method can achieve the workpieces’ recognizing, positioning and assembly accurately, meanwhile has high computational efficiency and strong robustness. As shown in the experiment results, for the recognition and positioning accuracy, the translation error of the proposed method is less than 0.2mm, and the rotation error is less than 0.2°; for the robot assembly accuracy, the translation error of the proposed method is less than 0.8mm and the rotation error is less than 0.4°. Therefore, the proposed method can successfully satisfy the requirement of robot assembly tasks.