A method of edge milling for industrial robots guided by binocular vision is proposed

which is suitable for machining workpieces composed of 3D planes. Firstly

the Zhang Zhengyou’s calibration method and the Tsai-Lenz hand-eye calibration method are used to calculate the internal/external parameters of binocular cameras and the hand-eye transition matrix between binocular cameras and an industrial robot. Then edges of workpieces are detected by binary processing of the binocular images. And corresponding relationships of points sampled from edges in the binocular images are established based on some preset ArUco markers. Finally

the 3D points on edges of workpieces are calculated according to the hand-eye transition matrix and the matched edge points. And robot milling trajectories are generated and sent to the robot controller so that the robot can execute edge milling processes. In the experiments of milling circular and straight edges

the results show that smooth edges are produced by our method and the maximum edge error is less than 0.5mm.