基于机器人运动模式的手眼标定初值估计法  

The Initial Estimates in Robotic Hand-Eye Calibration Based on Robot Movement Model

作　　者：刘宏建[1] 石繁槐[1] 刘允才[1] 罗毅[1] 

机构地区：[1]上海交通大学图像处理与模式识别研究所,上海200030

出　　处：《上海交通大学学报》2005年第3期396-399,共4页Journal of Shanghai Jiaotong University

摘　　要：针对目前机器人手眼标定的方法大多需要求解手眼标定的基本方程,提出了一种新的用于机器人手眼标定中初值估计的方法.该方法在非线性优化过程中需要对初值进行估计,不同的初值可得到不同的结果,传统技术主要通过经验选取初值.本文利用机器人自身的运动模式,直接确定初值,大大减小了经验选择给结果带来的偏差.与经验选择初值的方法进行对比,结果表明,本方法得到了明显的改善.Considering that in most hand-eye calibration methods, the basic equation needs to be solved, a new initial estimates method for robotic hand-eye calibration was proposed. Initial values are needed in optimizing the process of solving the equation, while in existing method, initial values are selected depending on experience. The value is determined directly by considering the robot movement model, which can (effectively) avoid the errors introduced by experience. Compared with classical methods, the proposed method shows a better improvement.

关 键 词：计算机视觉 机器人 手眼标定 初值估计 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]