A blending crossover differential evolution algorithm is proposed to increase the precision of camera-space manipula-tion (CSM) system. In this approach, six view parameters and flattening parameter are assembled into a single parameter of blending crossover differential evolution;the positioning precision of camera-space manipulation is set to be a fitness function.The CSM system can obtain the optimal parameter combination by evolutionary iteration.Experimental results of a virtual robot system show the robot positioning precision is improved by blending crossover differential evolution algorithm.%为了提高相机空间操控(CSM)系统的预测精度,提出一种基于混合交叉操作的差分进化算法。该方法将CSM系统的视觉参数初值和平化距离参数Z o进行组合作为混合交叉差分进化算法的个体,以CSM系统对目标点位置的预测精度作为个体的适应度函数,通过进化迭代获得最优的参数组合。使用了实际机器人视觉系统获取的数据进行实验,结果表明使用优化后的参数组合可以提高系统的预测精度。
展开▼