基于图像分割的置信传播立体匹配算法研究 - 红外技术.pdf

第33 卷 第 3 期 红 外 技 术 Vol.33 No.3 2011 年 3 月 Infrared Technology Mar. 2011 基于图像分割的置信传播立体匹配算法研究 李彬彬，王敬东，李 鹏 （南京航空航天大学自动化学院，江苏南京 210016 ） 摘要：在采用全局优化的立体匹配时，利用传统的基于像素点的置信传播算法进行消息传输存在两个 不足：计算量大和单个像素点容易导致误差。利用图像分割思想来弥补这两个不足，提出基于图像分 割的置信传播算法，利用分割区域代替像素点，视差平面模板集代替视差有哪些信誉好的足球投注网站空间范围，消息传输转 换为在分割区域间的传输，一定程度上降低了算法的计算量，同时利用图像分割的约束提高了算法对 遮挡区域和缺少纹理区域的处理能力。实验结果表明，算法相对于基于像素点的置信传播算法计算量 减小，匹配精度高，对缺少纹理区域和遮挡区域的处理更有优势，得到了较准确的稠密视差图。 关键词：立体匹配；置信传播；图像分割；视差平面模板 中图分类号：TP391 文献标识码：A 文章编号：1001-8891(2011)03-0167-06 Research of Stereo Matching Using Belief Propagation Based on Image Segmentation LI Bin-bin ，WANG Jing-dong ，LI Peng (College of Automation Engineering, Nanjing University of Aeronautics Astronautics, Nanjing 210016, China) Abstract ：The traditional stereo matching methods using belief propagation algorithm based on pixel have two deficiencies: high computation complexity and errors easily induced by single pixel. To make up for the two deficiencies, this paper proposes a novel belief propagation algorithm based on image segmentation, via using segmented regions instead of pixels and disparity plane set instead of disparity search space. Instead of assigning a disparity value to each pixel, a disparity plane is assigned to each segment. The amount of calculation is reduced and the capability to deal with textureless regions and occlusions is improved with segment constraint. Experimental results show that our algorithm has lower computation complexity, higher matching accuracy and better performance for processing in textureless regions and occluded regions than other belief pr