一种基于级联滤波的手眼在线标定方法  

An online hand-eye calibration approach based on cascaded filter

作　　者：陈天飞[1] 汪洋[1] 陈余庆[1] 吴翔[2] 马孜[1] 

机构地区：[1]大连海事大学自动化研究中心,大连116026 [2]河南工业大学电气工程学院,郑州450001

出　　处：《东南大学学报（自然科学版）》2013年第A01期138-142,共5页Journal of Southeast University：Natural Science Edition

基　　金：国家重大科技专项资助项目(2009ZX04001-021);国家自然科学基金资助项目(61203082);中央高校基本科研业务费专项资金资助项目(2011QN122)

摘　　要：在机器人末端关节处安装线结构光传感器可实现物体表面的非接触测量,针对该结构光测量机器人系统中的手眼标定问题,提出了一种基于级联滤波的在线标定方法.与以往先采集数据然后标定的离线方式不同,该方法在采集数据的同时可对手眼关系进行递归估计.首先对齐次标定方程AX=XB解耦展开,将展开后的子方程改写成状态空间模型,完成手眼关系级联滤波模型的建立.然后根据级联模型中子模型属性的不同,选择最适合的滤波方法,应用粒子滤波迭代估计手眼关系的旋转部分,用卡尔曼滤波实现平移向量的递推估计.实验表明了该方法的有效性,满足实际工程应用要求.A robot can be used for noncontact measurements when a line structured light sensor is fixed to its end-link.An online calibration approach based on cascaded filter is proposed for handeye problem in the robot-based structured light measuring system.Unlike previous offline methods w hich compute hand-eye relationship after data collection,the proposed approach can estimate handeye parameters recursively w hile acquiring the calibration data.First,the homogeneous matrix AX = XB is decoupled,and the expanded sub-equations can be rew ritten as a state space model.Then the cascaded filter model of the hand-eye relationship is established.According to the properties of submodel in the cascaded filter model,the most suitable filter algorithm is chosen.The particle filter algorithm is adopted to estimate the rotation part,and the Kalman filter algorithm is used to estimate translation vector.Finally,a real case study demonstrates that this method is effective and suitable for practical engineering application.

关 键 词：手眼标定 粒子滤波 卡尔曼滤波 级联滤波模型 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]