摘要
为了实现视觉获取目标物立体信息,以立体测量原理为基础,采用Matlab和Opencv相结合的方法,设计了基于双目立体视觉的目标物定位系统。该系统采用棋盘标定法,利用Matlab完成双目相机标定,采用BM(Block Matching)立体匹配算法在VS2010环境下配合Opencv3.0.0库,完成左右图像对的匹配,生成视差图。通过将二维空间点重投影至三维空间,便可求得目标点的空间三维坐标,最终实现通过鼠标点选后输出选定点的空间坐标。实验表明,该系统拥有较高的测量精度,尤其在150cm距离下可获得最佳的测量精度。
To get stereo information of the object by computer vision,this paper taked the approach of stereo vision measuring principle and designed a location system based on binocular stereo vision.This system was accomplished by using Matlab and Opencv.The binocular camera calibration was accomplished by using checkerboard calibration method.The system accomplished stereo matching by using block matching algorithm based on VS2010 and Opencv3.0.0.The system can generate the disparity map and get the 3 Dcoordinates of the object by reprojecting the point in 2 Dspace to 3 Dspace.Finally the system can output the coordinates which you choose by clicking on a mouse.The results shows that this system can work with a high precision and achive the highest precision at the distance of 150 cm.
出处
《软件导刊》
2018年第2期198-201,共4页
Software Guide
关键词
立体视觉
目标定位
相机标定
立体匹配
重投影
stereo vision
target location
camera calibration
stereo matching
reprojection
