基于快速立体匹配的基本矩阵估计方法  

Fundamental Matrix Estimation Method Based on Fast Stereo Matching

作　　者：刘丁[1] 赵豆[1] 杨延西[1] 

机构地区：[1]西安理工大学自动化与信息工程学院,陕西西安710048

出　　处：《西安理工大学学报》2009年第1期1-7,共7页Journal of Xi'an University of Technology

基　　金：国家自然科学基金资助项目(60675048;60805020)

摘　　要：借鉴Harris角点检测时角点自身的特点,提出了基于角特征与灰度相关并参考视差梯度约束的角点匹配方法,利用角特征值比值的约束条件,减少了大量的候选匹配点,提高了立体匹配的速度,实现了快速、准确的立体图像对的角点匹配。在此基础上,进一步采用Hartley提出的改进八点算法,获得了高精度的基本矩阵。实验结果表明该方法具有较高的准确性和实时性,是一种快速有效的立体匹配基本矩阵估计方法。The comer characteristic detected by Harris algorithm is used for reference, and comer matching method based on comer characteristic and gray correlation by selectively adopting gradient of disparity constraint is proposed. An amount of alternative comers are reduced by using the ratio between comer eigenvalue, so as to improve the stereo matching velocity. As a result, a fast and exact comer matching method for stereo image couples is realized, on the basis of which, normalized eight-point algorithm proposed by Hartley is used to estimate the high-precision fundamental matrix. Many experiment results show that the method in this paper is of high accuracy and rea-time as well as a high-precision fundamental matrix estimation method based on fast stereo matching.

关 键 词：HARRIS角点检测 角特征值 灰度相关 立体匹配 基本矩阵 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]