Research On Stereo Matching Algorithm Based On Image Segmentation

Stereo version is a hot topic in the field of Computer Version, and stereo matching is the foundation and key technology in stereo version. In the research of the conversion of two visual angles to multiple visual angles in naked eye stereo display,we found the problem that traditional stereo matching algorithms cannot satisfy the requirement of both speed and preciseness. This paper focus on the issue of speed and preciseness in stereo matching and image segmentation.The main work of this paper can be listed as follows:Firstly,this paper introduces key technologies and algorithms in stereo version. And it particularly describes stereo matching, including the fundamental and restrictions, as well as some matching algorithms. Furthermore, analyze the theory of image segmentation,including its base model and algorithms.Secondly,we propose a kind of stereo matching algorithm,based on image segmentation,combining stereo matching and image segmentation. In terms of image segmentation,we use mean-shift algorithm and relevance between frames to improve the segmentation speed.of a vedio stream. And the matching process is based on the segmental areasAt last,we implement the stereo matching algorithm based on image segmentation proposed above. And experiment presents that the model proposed by this paper have good effect and satisfys the requirements.