一种改进的机器人测距定位摄像机标定方法  

An improved camera calibration method of robot distance measuring

作　　者：顾冬华[1] 王宏[1] 孙冬[1] 

机构地区：[1]郑州轻工业学院工程训练中心,河南郑州450002

出　　处：《郑州轻工业学院学报（自然科学版）》2015年第5期121-123,共3页Journal of Zhengzhou University of Light Industry:Natural Science

摘　　要：针对基于主动视觉的摄像机标定方法对摄像机运动形式要求过高且不易实现的问题,提出一种只需摄像机进行两次纯平移运动的简化的标定方法.该方法由两对匹配点对线性求解基本矩阵F,并由基本矩阵的内在关系得出摄像机内参数的线性解,进而求得4个内参数.仿真实验结果表明,在满足一定条件的前提下,该方法可以将仿真误差控制在技术标准许可的范围之内,可满足机器人测距定位的精度要求,且减少计算量,提高机器人测距的实时性.Aiming at the problem that camera calibration method based on active vision was too high in requirement in the form of camera motion,and difficult to realize,a simplified method of camera calibration was proposed which needed only 2 times pure translation motion. In this method,the basic matrix F was obtained by 2 pairs of matching points,and the linear solution of the internal parameters of the camera was obtained by the intrinsic relationship of the basic matrix,then the 4 internal parameters were obtained. The simulation experiment results showed that this method could be controlled within a certain range of simulation error under certain conditions,which could meet the requirement of the accuracy of the robot location,and reduce the computation time and improve the real-time performance of the robot.

关 键 词：摄像机标定 机器人测距定位 基本矩阵 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]