Resarch The Stere Matching Methods Based On The Graph Cuts And The OpenCV

The technique of stere matching is the key subject in computer visionfield,which is widely applied in precise industrial measurement,medicaldevice,object recognition,virtual reality,auto navigation,automatization ofproduction,military affair and many other fields.At present,the stere matchingmethods based on the Grap Cuts is the hot spot in the field of computer vision,andthe OpenCV library vision which develop at the computer vision at the latest.So,it isvery important to research to the stereo matching algorithm based on the Graph Cutsand the OpenCV library vision in this paper.Concretely speaking,it has mainly the following innovative at this paper:(1) We delves into the the stereo matching methods based on the OpenCVlibrary,How to compute the epipolar lines for the OpenCV function,the epipolargeometry,the essential and fundamental matrices,the stereo calibration,stereorectification,rectification map and the stereo correspondence methods in this paper.(2) As the complexity of the global match algorithm is too high, a new fast androbust stereo matching algorithm based on graph-cut is proposed. On the one hand,according to the local stereo matching algorithm, the nodes and edges in thegraph-cut are reduced. It is unnecessary to traverse all parallax when getting theglobal optimization, so the match time is reduced. On the other hand, the2D linearinterpolation algorithm is used for calculating the sub-pixels, so the precision androbustness are improved. The experiments demonstrate the algorithm is fast, robustand accurate.(3) For the stereo matching problem in the textureless, occluded and depthdiscontinuity regions, a new stereo matching algorithms that based on the adaptivesupport-weight of Graph Cuts is proposed at this paper, which can reduce thematching error in the depth discontinuity and textureless regions by the singleadaptive support-weight matching methods. The occlusion and smoothness penaltyis considered by building the energy function. The experimental results show that theproposed algorithm can achieve more precise and reliability matching.