光学 精密工程, 2018, 26 (2): 461, 网络出版: 2018-03-21   

快速鲁棒的基础矩阵估计

A fast and robust method for fundamental matrix estimation
作者单位
1 中国科学院 光电技术研究所,四川 成都 610209
2 中国科学院大学, 北京 100190
摘要
针对基础矩阵估计过程中因受野值影响导致估计精度下降和稳定性不高等问题,本文提出了一种新的快速鲁棒的基础矩阵估计方法。该方法首先将野值去除融入到计算基础矩阵的过程中,而不再将它作为一个独立的处理步骤。通过迭代将潜在的错误对应点剔除,从而实现基础矩阵的稳定估计。然后,在每次迭代过程中,采用对极几何误差准则来识别野值,同时获得基础矩阵的估计结果。该迭代过程收敛较快,即使存在大量匹配野值的情况下,计算值也会很快趋于稳定。仿真和实际实验结果一致表明:所提出的算法在保证类似估计精度的同时还在计算效率方面有极大地提升,相比较快的M估计法有30%以上的速度提升,而相比于估计精度较优的MAPSAC算法甚至达到4倍以上。
Abstract
In this paper, a new fast and robust fundamental matrix estimation method was proposed to solve the problem that the estimation of fundamental matrix leads to lower estimation accuracy and lower stability due to outliers. The method removed outliers into the computation of the fundamental matrix instead of taking it as an independent processing step. The potential error corresponding points were eliminated by iteration to achieve the stable estimation of the fundamental matrix. Then, the epipolar geometry error criterion was used to identify outliers and the estimation results of the fundamental matrix were obtained during each iteration. The iterative process could converge quickly, even if a large number of matched outliers were present, the calculated values would soon become stable. The results of simulation and actual experimental show that the proposed algorithm improves the estimation accuracy greatly, and also ensures similar calculation efficiency at the same time. Compared with the method of M-estimator, it has more than 30% speed improvement, and compared with the MAP-SAC algorithm with higher estimation accuracy, it even achieves more than 4 times.

颜坤, 刘恩海, 赵汝进, 田宏, 张壮. 快速鲁棒的基础矩阵估计[J]. 光学 精密工程, 2018, 26(2): 461. YAN Kun, LIU En-hai, ZHAO Ru-jin, TIAN Hong, ZHANG Zhuang. A fast and robust method for fundamental matrix estimation[J]. Optics and Precision Engineering, 2018, 26(2): 461.

本文已被 1 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!