Abstract：Aiming at the limitations of the traditional non-parametric transformation, a stereo matching algorithm is proposed. One matching cost based on support region with arbitrary shape and size is obtained by using color segmentation information of the reference image, and another matching cost of weighted non-parametric transform in similar color region is obtained, then joint matching cost is obtained by the fusion of the both matching cost, and the dense disparity map is realized through the local optimization method. The experiment results show that the algorithm can improve the accuracy of the low texture, borders and occlusion regions. At the same time, it is robust against the amplitude distortion.
［1］康新,何小元,冯毅.一种新的模板匹配算法及其在三维形貌测量中的应用［J］.计量学报, 2002,23(2):90-93.
［2］Scharstein D, Szeliski R. A taxonomy and evaluation of dense two-frame stereo correspondence algorithms［J］. International Journal of Computer Vision, 2002, 47(1-3): 7-42.
［3］Veksler O. Stereo correspondence by dynamic programming on a tree［C］//IEEE Computer society Conference on Computer Vision and Pattern Recognition. Washington,USA, 2005:384-390.
［4］于英,杨靖宇,张永生,等.一种适应无人机平台的快速立体匹配方法［J］.计量学报,2014,35(2):143-147.
［5］Yang Q, Wang L, Ahuja N. A constant-space belief propagation algorithm for stereo matching［C］//IEEE Conference on Computer Vision and Pattern Recognition. San Francisco,USA,2010:1458-1465.
［6］Cheng F, Zhang H, Yuan D, et al. Stereo matching by using the global edge constraint［J］. Neurocomputing, 2014, 131(7): 217-226.
［7］Yoon K J, Kweon I S. Adaptive support-weight approach for correspondence search［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2006, 28(4): 650-656.
［8］Nalpantidis L, Gasteratos A. Biologically and psychophysically inspired adaptive support weights algorithm for stereo correspondence［J］. Robotics and Autonomous Systems, 2010, 58(5): 457-464.
［9］Zabih R, Woodfill J. Nonparametric local transforms for computing visual correspondence［C］//Third European Conference on Computer Vision, Stockholm, Sweden, 1994 :151-158.
［10］Kelly Gazolli, Evandro Salles. Exploring neighborhood and spatial information for improving scene classification［J］. Pattern Recognition Letters, 2014, 46(3): 83-88.
［11］Ma L, Li J, Ma J, et al. A modified census transform based on the neighborhood information for stereo matching algorithm［C］//The Seventh International Conference on Image and Graphics. Qingdao,China, 2013:533-538.
2017, Vol. 38 
