一种月球车的环境重建方法  

A Method of Environment Reconstruction for Lunar Rover

作　　者：戴宪彪[1] 王亮[1] 居鹤华[1] 

机构地区：[1]北京工业大学电子信息与控制工程学院,北京100124

出　　处：《计算机测量与控制》2011年第7期1699-1701,1704,共4页Computer Measurement &Control

基　　金：国家863高科技资助项目(2008AA0085)

摘　　要：根据对月球车双目立体视觉系统及其工作环境的分析,提出了一种基于相机离线标定和区域生长的稠密匹配相结合的月球车环境重建方法,通过相机离线标定、极线校正和基于区域生长的立体匹配,得到月球车漫游环境的重构地图;首先,与常用的相机标定方法相比,基于OpenCV的相机离线标定在得到可靠的相机参数的同时,减少了三维重建的时间,提高了实时性;其次,通过极线校正和改进的区域生长立体匹配方法可以快速的得到稠密的月球车的环境深度图;实验结果证明了该方法的有效性。A 3-D reconstruction method is presented based on the analysis of the stereo vision system and working surroundings of Lunar Rover,which includes camera offline calibration and region-growth match.The depth map surrounding Lunar Rover is generated by camera offline calibration,epipolar correction and dense match based on region-growth algorithm.Firstly,camera offline calibration based on OpenCV reduce the construction time compare to common method of camera calibration,at the same time it can gain the reliable camera parameters;Then,we could fastly obtain dense depth map using epipolar correction and improved dense match based on region-growth algorithm.The experiments shows the validity of the proposed method.

关 键 词：月球车 双目立体视觉 三维重建 立体匹配 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]