基于主动视觉的测量装置精度校准方法

来源期刊:中南大学学报(自然科学版)2014年第4期

论文作者:王宪 谭建平 陈国强 成小乐

文章页码:1071 - 1077

关键词:在线监测;计算机视觉;精度校准;主动视觉

Key words:online monitoring; computer vision; accuracy calibration; active vision

摘    要:针对工业在线监测对测量装置精度及稳定性的要求,对视觉测量装置精度校准及维护方法进行研究。分析透视原理下引入图像畸变的测量装置成像模型,推导测量装置相对基准激光做直线运动时光斑在畸变图像中的轨迹方程,提出一种基于主动视觉的测量装置精度校准方法:首先,通过控制测量装置做二维平移运动得到校准所需的特征点信息;接着,利用光斑的畸变轨迹方程求解测量图像的畸变系数并对图像进行畸变校正;最后,求解图像平面到被测平面的单应性矩阵从而完成精度校准。对特征点的布置方式进行优化。研究结果表明:校准后的定位测量峰值误差为0.373像素,均方根误差为0.178像素。该方法精度较高且易于实施,能满足工业在线检测的要求。

Abstract: The method for the accuracy calibration and maintenance of vision measuring device was studied to meet the requirement of the accuracy and stability of the measurement device in industrial online monitoring. An imaging model to which image distortion was introduced in the principle of perspective was analyzed. The trajectory equation of spots moving in the distortion image when the measuring device was in linear motion in relation to the laser was deduced. The method for accuracy calibration of active vision-based measurement device was proposed. The procedures were as follows: Firstly, gathering the information on feature points necessary for calibration from the in-plane translational motion of the measurement device under control; secondly, obtaining the distortion coefficient of measurement image by the spot trajectory equation and correcting the image distortion; and at last, getting the homography matrix between the image plane and the measured plane to complete the accuracy calibration. The layout of the feature points was optimized. The experimental results indicate that the peak value error of centering measurement is 0.373 pixel, and the root-mean-square error is 0.178 pixel. The method is easily applicable and accurate enough for industrial online monitoring.

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号