Dual quaternion hand-eye calibration algorithm for hunter-prey optimization based on twice opposition-learning and random differential variation

Abstract

In order to solve the problem of noise interference in camera calibration and robot forward kinematics solution, this study proposes a dual quaternion hand-eye calibration algorithm based on twice opposition-learning and random differential variation (ODHPO). The hand-eye calibration equation was innovatively rewritten in the form of dual quaternion, and the F-norm minimization model of the rotation and translation error function was constructed for optimization. At the same time, the penalty function method is introduced to effectively transform the constrained optimization problem in hand-eye calibration into an unconstrained problem, which is further solved by the ODHPO algorithm for global optimization. Compared with the singular value decomposition algorithm based on the traditional dual quaternion method, the ODHPO algorithm performs better in global optimization ability and convergence stability. Through numerical simulation and real robot hand-eye calibration experiments, it is proved that the proposed algorithm is superior to the traditional dual quaternion (CDQ) algorithm, the classical algorithms Tsai method and Navy method in terms of solution accuracy, sensitivity to the number of pose transformations and stability, demonstrating its potential for application and practical significance in robot vision system.