基于避障的移动机器人立体视觉算法研究-精密仪器与机械专业论文.docxVIP

中国科学技术大学硕士论文 中国科学技术大学硕士论文 英文摘要 Abstract Nowadays，autonomous mobile robot can be applied to military areas， manufacturing，and everyday living etc．As a result，it is an active and promising area and gets much more world wide attention．However，as a key part of mobile robot techniques，many techniques of robot vision are still iiTlrnature．Robot vision． especially stereo vision．is involved in several different projects in my 1ab．To fulfilI the requirement of也ese projects，we foCUS our research on stereo vision techniques． especially camera calibration，stereo matching and obstacle detection etc．The main work and contribution are following： 1)Describing a new flexible method of camera calibration which has wonderful performance and is also easy to conduct．This technique can be applied to any other stereo vision system without any modification． 21 Wj propose a new stereo matching algorithm in this paper to improve the performance of stereo vision in environments presenting repetitive patterns or regions with relatively weak texture．This algorithm alleviates the correspondences problem for pixels that can not be reliably matched by some other stereo algorithms． 3)Based on the technique of camera calibration，we rectify the images to a)deal 、vith mdial distortion and协all corresponding pixels exist at the same scan line． This technique not only improves the efficiency but also makes it easier to find corresponding pixel in matching process． 4)Our stereo vision system does not need precise 3-D scene reconstruction．It reconstructs a part ofthe scene，especially the obstacle，at a relatively imprecise level．In this way,we can on the one hand decrease the time spending reconstruction．On t11e other hand．we can still obtain enough obstacle informati011 without detecting the details of the obstacle．It cart be seen for our experiment that this algorithm works very