An Improved Algorithm of the Hand-Eye Vision Positioning System Based on the Constant Rotation Matrix

The implementation of object 3D position has been greatly simplified by robot hand-eye vision positioning system which is based on the constant rotation matrix. The main idea is to simplify the complicated calibration procession of robot visual control system by maintaining the constant rotation matrix that maps hand coordinate to robot coordinate. Given the camera light axes being vertical with the working table, the calibration of parameter matrices is carried out by moving the hand to different positions and taking pictures for a put object. In this paper, the authors present a new method for calibrating parameter matrices, which simplifies the process of object 3D position and improves precisions. The control system of MOTOMAN-SV3X robot has realized this new method, and examples are given to illustrate the efficiency of the algorithm