摘要
深入研究了机器人手眼视觉系统的立体定位问题。首先,重新将手眼问题公式化,得到一个新的系统模型;然后,在此基础上提出了一种实用有效的标定方法。其核心思想是直接将图像坐标映射到机器人基坐标系中,把系统参数作为一个整体来获取,而不必分别计算摄像机内部的每一个参数。与原方法相比,本方法可随意改变末端姿态定位,即定位时摄像机对目标取像的姿态不受任何约束。实验表明,该方法操作方便、实现简单、定位精度高。这一方法的提出克服了原方法的局限性,大大推广了手眼视觉系统的应用范围。
In this paper,the hand-eye problem is reformulated and a new system model is obtained,by which an effective system calibration method is presented.The key is directly mapping the camera coordinate to the robot coordinate,and the system parameters can be obtained as a whole without considering every one thereln.Compared with the original method,there is not any restriction to the system positioning,that is to say,the camera can take images at any pose and the system needn＇t to be recalibrated.Experiments show that this algorithm is quick and convenient to be applied; the experimental procedure is simple and positioning results are precise.This method overcomes the limitation of the original one and extends the application range of handeye vision system.
出处
《计算机工程与应用》
CSCD
北大核心
2007年第24期235-238,共4页
Computer Engineering and Applications
关键词
机器人手眼视觉系统标定摄像机标定
robot
Hand-eye Vision
System Calibration
camera calibration
