A calibration method combining hand-eye calibration and TCP calibration

Abstract Robots' traditional Tool Center Point (TCP) calibration and hand-eye calibration are implemented independently, which is likely to generate large cumulative errors. In this paper, a collaborative calibration method integrating TCP calibration and hand-eye calibration is put forward. The nine-point method and the planar TCP calibration method are employed to acquire the initial matrix for hand-eye calibration and the initial TCP offset, and to set up the coordinate equation of an arbitrary point in space under the base coordinate system. Through the Lie algebra linear dimensionality reduction method, the TCP value, rotation angle, camera height, and the error compensation amount of the hand-eye matrix are obtained, thus getting the precise values of hand-eye calibration and TCP calibration. Through experimental verification of the surface, the collaborative calibration algorithm adopted in this paper reduces the position error and attitude error, improves the calibration accuracy of the robot, and can meet the needs of more working scenarios.