摘要： 为提高基本矩阵估计精度，提出一种改进的随机抽样最大似然估计算法。根据对极距离选择质量较好的原始数据，采用随机抽样一致性方法进行抽样，选择内点数最多的基本矩阵检验原始数据，剔除误差大的匹配点，结合约束条件对匹配集进行检验，以提高匹配集精度。实验结果表明，该算法的估计精度较高，稳定性较好。

关键词: 基本矩阵, 随机抽样一致性算法, 随机抽样最大似然估计算法, Sampson误差, 初始值

Abstract: In order to improve the accuracy of fundamental matrix estimation, an improved Maximum Likelihood Estimation by Sample and Consensus(MLESAC) algorithm is proposed. According to the distance between the matching points and the corresponding epipolar lines, the superior correspondences are chosen, random sample consensus is adopted to sample the superior correspondences, the fundamental matrix with the largest number of inliers is chosen to examine all the corresponding points and eliminate mismatches. It detects matching points set according to the epipoplar geometry and adding constraints, and the accuracy of matching set is improved. Experimental results show that the accuracy of this algorithm is improved, and the stability is better.

Key words: fundamental matrix, Random Sample Consensus(RANSAC) algorithm, Maximum Likelihood Estimation by Sample and Consensus (MLESAC) algorithm, Sampson error, original value

中图分类号: 

• TN911.73