基于ICP算法的苹果采摘机器人手眼标定方法

苹果采摘机器人手眼协同作业过程中手眼关系的精度是影响机器人作业效率和作业精度的关键因素之一.针对传统手眼标定方法求解精度低、线性化求解时解耦精度丢失和非线性求解时累积误差较大等问题,提出一种基于ICP算法的手眼标定方法.以Tsai〖CD*2〗Lenz法的求解结果作为ICP匹配算法初值,通过确定ICP匹配算法中误差阈值和迭代次数最优组合得到精确的手眼转换关系.试验结果表明,本文方法避免了求解时精度丢失和误差积累等问题,与Tsai〖CD*2〗Lenz算法相比,旋转精度提高了13.6%,平移精度提高12.47%.

Hand-eye Calibration Method for Apple Picking Robot Based on ICP Algorithm

In the process of hand-eye collaboration of apple picking robot, the accuracy of hand-eye relationship is one of the key factors affecting the efficiency and accuracy of robot operation. In order to solve the problem of low accuracy of traditional hand-eye calibration methods, linear and nonlinear solving methods based on mathematical tools such as quaternions are mostly adopted. In this paper, a hand-eye calibration method based on ICP algorithm is proposed to solve the problems of loss of decoupling accuracy during current linearization solution and large cumulative error during nonlinear solution. This method uses the solution result of Tsai-Lenz as the initial value of the ICP matching algorithm, and substitutes it for iterative calculation, and obtains the accurate hand-eye conversion relationship by determining the optimal combination of error threshold and number of iterations in the ICP matching algorithm. Experimental results show that the hand-eye calibration optimization algorithm based on ICP not only avoids the problems of accuracy loss and large error during solving, but also improves the rotation accuracy by 13.6% and the translation accuracy by 12.47% compared with Tsai-Lenz.