基于场景层次的遮挡建模与立体匹配  

Hierarchy-Based Occlusion Modeling and Stereo Matching

作　　者：李旭东[1] 张凯[2] 汪国平[2] 

机构地区：[1]天津大学计算机科学与技术学院,天津300072 [2]北京大学信息科学技术学院,北京100871

出　　处：《计算机辅助设计与图形学学报》2008年第10期1358-1365,共8页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家“九七三”重点基础研究发展规划项目(2004CB719403);国家“八六三”高技术研究发展计划(2004AA115120)

摘　　要：通过在扫描线和图像分割2个层面上建立基于场景层次的遮挡模型,实现了一种全新的从前景到后景分层处理的立体匹配算法.以像素灰度值为依据进行匹配时,离摄像机最近的物体因为不受遮挡的影响,总是最容易匹配成功;稍远一些的景物可以利用已经计算出来的前景建立局部的遮挡模型指导匹配.这样逐层递推,可求得场景中所有物体的视差.图像分割技术将处理单位由像素提升到了图像块,既提高了算法效率,也降低了视差映射图的不连续性.使用标准数据集的测试结果表明,文中算法精度高、强壮性好,在无纹理区域、深度变化边界区域和遮挡区域都达到了比较高的识别率.A new occlusion model is presented. The occlusion model based stereo matching algorithm using scene hierarchical structure is constructed on two levels： image scan-line and segment. In particular, the paper highlights a previously overlooked geometric can be easily detected by intensity-based cost function and the rarer fact： the most foreground objects objects can be matched using local occlusion model constructed by the former recognized objects. Then the scene structure is achieved from foreground to background. Image segmentation is adopted to increase the algorithm＇s efficiency and to decrease the discontinuity of disparity map. Experimental results demonstrate that the proposed algorithm is among the state-of-the-art stereo algorithms on various datasets. Furthermore, better performance is achieved in the conventionally difficult areas such as texture-less regions, disparity discontinuous boundaries and occluded portions.

关 键 词：立体匹配 视差层次 遮挡模型 代价函数 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]