A Positioning System Based on Monocular Vision for Industrial Robots

Abstract

In order to make the industrial robot catch the target object accurately, an automatic positioning system based on monocular vision is proposed in this paper. Specifically, efficient localization algorithms are applied to get the position information of the target object. Besides, in consideration of further application requirements, an additional recognition algorithm based on 2-1 Hough transform is adopted to recognize the welding points on the accumulator. Moreover, a pose estimation algorithm is also used to obtain the posture information of the object. Then the information is converted to the robot's basal coordinate and transmitted to the robot motion controller by a serial port. Finally, a trajectory planning algorithm based on circular interpolation is applied for the 6-DOF industrial robot to capture the objects. The experimental results demonstrate that the proposed methods can recognize and acquire the pose of the target object accurately, so as to guide the industrial robot to catch the object reliably. The proposed positioning system can well meet the requirements of actual industrial productions.