Abstract:Hand-eye calibration determines the nonlinear mapping relationship between robot base coordinate system and camera coordinate system, and it plays an important role in visual servo. Aimed at the issue of hand-eye calibration in visual servo control system, based on robot toolbox and neural network toolbox, under the environment of MATLAB/simulink, the error back propagation (BP) neural network algorithm and radial basis function (RBF) neural network algorithm were used to simulate the mapping relationship between 6-DOF sorting robot and monocular camera. The accuracy of the two algorithms was analyzed through the simulation results. In addition, the hand-eye calibration of the manipulator was carried out by using BP neural network and Zhang’s method under the same experimental conditions. The same group of random sample points were grabbed in the actual workspace of the manipulator, and the grasping accuracy of the random sample points was compared. The simulation and experimental results showed that the calibration accuracy of BP neural network was better than that of RBF neural network and Zhang’s calibration method, and could improve the accuracy of hand-eye calibration in practical application.