机器人
機器人
궤기인
ROBOT
2010年
1期
104-108
,共5页
立体匹配%图割%简化网格图%最小割最大流
立體匹配%圖割%簡化網格圖%最小割最大流
입체필배%도할%간화망격도%최소할최대류
stereo matching%graph cut%reduced graph%min-cutmax-flow
针对基于图割法的立体匹配算法耗时太长的问题,提出了一种基于简化网格图的立体匹配算法.算法通过区域匹配算法得到每个像素的初始视差值,然后只保留完整网格图的部分可能的视差值,去除其余大部分的节点和边缘,建立简化的网格图.该方法大大缩减了网格图的容量,缩短匹配所用时间,并且能够选用更大的视差范围.实验证明,该算法能够得到比较理想的视差图,而且人大缩短屯体匹配所用时间.
針對基于圖割法的立體匹配算法耗時太長的問題,提齣瞭一種基于簡化網格圖的立體匹配算法.算法通過區域匹配算法得到每箇像素的初始視差值,然後隻保留完整網格圖的部分可能的視差值,去除其餘大部分的節點和邊緣,建立簡化的網格圖.該方法大大縮減瞭網格圖的容量,縮短匹配所用時間,併且能夠選用更大的視差範圍.實驗證明,該算法能夠得到比較理想的視差圖,而且人大縮短屯體匹配所用時間.
침대기우도할법적입체필배산법모시태장적문제,제출료일충기우간화망격도적입체필배산법.산법통과구역필배산법득도매개상소적초시시차치,연후지보류완정망격도적부분가능적시차치,거제기여대부분적절점화변연,건립간화적망격도.해방법대대축감료망격도적용량,축단필배소용시간,병차능구선용경대적시차범위.실험증명,해산법능구득도비교이상적시차도,이차인대축단둔체필배소용시간.
For the problem that stereo matching methods based on graph cuts are time consuming, this paper puts forward an improved stereo matching algorithm based on reduced graphs. First, the initial disparity for each pixel can be calculated by using local matching method. Then, we keep only some potential disparity values in the complete graph, and the reduced graph will contain a reduced number of vertices and edges. Therefore the graph capacity and execution time are decreased, and a wider disparity range is obtained. At last, it is proved by experiments that this algorithm can achieve a relatively ideal disparity map and save much time in stereo matching.