基于极线几何的机器人视觉伺服控制系统分析  

The Analysis of Robotic Visual Servo Control System Based on Epipolar Geometry

作　　者：杨逢瑜[1] 王其磊[1] 关红艳[1] 陈君辉[1] 王顺[1] 娄磊[1] 

机构地区：[1]兰州理工大学流体动力与控制学院,甘肃兰州730050

出　　处：《西华大学学报（自然科学版）》2009年第1期14-16,20,共4页Journal of Xihua University:Natural Science Edition

摘　　要：为了更好地解决机器人视觉伺服控制系统中运行误差的问题,引入了一种基于图像的交互式几何建模方法。该方法利用自动方式标定模型的平行线、特征点,得到摄像机参数,进而计算出模型的几何信息,然后从图像中提取要素,提出一种最优控制运行方式,并且以此为基础提出了视觉伺服控制器的设计。仿真与试验结果表明,这种控制方法能够使图像上特征向量的误差收敛到零,机器人能够到达期望的位姿上。In order to solve the problem of function error for robotic visual servo control system, an interactive method of image- based modeling is presented. In this paper, the parallel lines and characteristic points of a model, which are gained from interactive in- put, are used to compute camera＇s parameters, and further model＇s geometric information. Texture information is also obtained from the images. An optimal control mode is proposed, and the design of visual servo controller is put forward based on the mode. The simula- tion and experiment results show that this control method can make the error of characteristic vectors of an image converge to zero, and the robot can reach the desired position and orientation.

关 键 词：极线几何 机器人 视觉伺服控制 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]