基于特征约束及区域相关的体视匹配方法

立体匹配是计算机视觉领域的一个关键问题,同时也是难点问题。为了得到准确的高密度视差图,通过对基于区域和基于特征的体视方法的讨论,综合两种方法的优点,提出了基于边缘特征约束及区域相关的立体匹配算法。该方法首先利用基于特征技术来得到边缘特征点,对边缘特征点再做灰度等区域相关匹配处理,然后在匹配的边缘特征点约束下,对非边缘特征点采用区域相关算法进行匹配,得到整体高密度视差图。这样既缩小了匹配搜索空间,又保证了匹配的可靠性。实验结果表明,该算法具有良好的效果和实用价值。

A Feature-Constrained and Area-Related Stereo Matching Algorithm

The stereo matching is a key problem of computer vision,and also is the difficult problem.In order to obtain precise matching and dense disparity map,a feature-constrained and area-related stereo matching algorithm is presented by way of synthesizing of feature-based and area-based process method.First.the authors obtain sparse feature points of edges using feature-based approach and get the matching information of them by using area-based algorithm.Then continue to fill in the disparity data of featureless points under the constraint of the acquired disparity map.In this way,a dense disparity map is showed.The method not only reduces the search difficulty but also ensures the validity of matching.The experimental results show this algorithm is effective and has practical value.