基于变异粒子群优化的在线手眼标定算法

Online hand-eye calibration algorithm based on variation particle swarm optimization

为满足机械臂手眼标定过程中不依赖人工提供固定机械臂路径的要求,提出一种基于变异粒子群优化的在线手眼标定算法.通过分析特征点在机械臂极坐标系与深度相机坐标系下的坐标转换关系,建立了基于深度相机的手眼模型;采用基于变异粒子群的在线优化算法,实时估计手眼关系参数;建立机械臂运动能量函数,给出了机械臂位姿补偿方法,满足特征点视野约束的同时保证运动能量最小.实验结果表明:所提方法能较为精确地估计手眼关系参数,并且能实时补偿机械臂位姿,实现自动手眼标定功能.

To satisfy the requirement of being independent of manually providing fixed manipulator path in the process of hand-eye calibration,an online hand-eye calibration algorithm based on the variation particle swarm optimization was proposed.The hand-eye model with depth camera was built by analyzing the coordinate conversion relation of feature points in the polar coordinate system of manipulator and the coordinate system of depth camera.The online algorithm based on the variation particle swarm optimization was applied to estimate the hand-eye parameters in real time.The motion energy function of manipulator was built and a compensation method of the manipulator position and posture was provided,which ensured the minimum motion energy while satisfying the visual field constraint of feature points.Experimental results show that the proposed method can accurately estimate the hand-eye parameters and compensate the position and posture of the manipulator in real time,realizing the automatic hand-eye calibration.