基于辅助匹配的1点RANSAC单目视觉导航算法

doi:10.3969/j.issn.1001-506X.2018.05.23

系统工程与电子技术 ›› 2018, Vol. 40 ›› Issue (5): 1109-1117.doi: 10.3969/j.issn.1001-506X.2018.05.23

基于辅助匹配的1点RANSAC单目视觉导航算法

齐乃新¹, 张胜修¹, 曹立佳^2,3, 杨小冈¹

1. 火箭军工程大学控制工程系, 陕西西安 710025; 2. 四川理工学院自动化与信息工程学院, 四川自贡 643000; 3. 人工智能四川省重点实验室, 四川自贡 643000

出版日期:2018-04-28 发布日期:2018-04-25

Monocular visual navigation method with 1-point RANSAC based on aided matching

QI Naixin¹, ZHANG Shengxiu¹, CAO Lijia^2,3, YANG Xiaogang¹

1. Department of Control Engineering, Rocket Force University of Engineering, Xi’an 710025, China; 2. College of Automation and Information Engineering, Sichuan University of Science & Engineering, Zigong 643000, China; 3. Artificial Intelligence Key Laboratory of Sichuan Province, Zigong 643000, China

Online:2018-04-28 Published:2018-04-25

摘要/Abstract

摘要：

针对1点随机抽样一致性(random sample consensus,RANSAC)单目视觉导航算法中的主动视觉匹配失效问题,提出了一种基于辅助匹配的1点RANSAC单目视觉导航算法。首先,该算法通过引入尺度不变特征变换(scale invariant feature transform,SIFT)算法完成特征匹配;其次,采用RANSAC算法解算基础矩阵和匹配点;最后,通过实验验证了算法的有效性。实验结果表明,该算法能够解决主动视觉匹配失效问题,提高1点RANSAC单目视觉导航算法的导航精度。SIFT辅助求解的有效匹配点精度在5个像素之内,航向角平均误差减小5.04°,俯仰角平均误差减小1.21°,滚动角平均误差减小3.03°。

Abstract:

In view that the active visual matching algorithm in the monocular visual navigation method with 1-point random sample consensus (RANSAC) has a matching failure problem, a monocular visual navigation method with 1-Point RANSAC based on aided matching is proposed. Firstly, the scale invariant feature transform (SIFT) algorithm is introduced to complete the feature matching. Secondly, the RANSAC algorithm is used to calculate the fundamental matrix and the matching points. Finally, the effectiveness of the proposed method is verified by experiments. Results show that the active visual matching failure problem is solved and the accuracy of the the 1-point RANSAC navigation method is improved. The accuracy of the matching points calculated by the SIFT algorithm is within 5 pixels. The mean errors of the camera’s course, pitch, and roll angles estimated by the proposed method are decreased by 5.04°, 1.21°, and 3.03°, respectively.

齐乃新, 张胜修, 曹立佳, 杨小冈. 基于辅助匹配的1点RANSAC单目视觉导航算法[J]. 系统工程与电子技术, 2018, 40(5): 1109-1117.

QI Naixin, ZHANG Shengxiu, CAO Lijia, YANG Xiaogang. Monocular visual navigation method with 1-point RANSAC based on aided matching[J]. Systems Engineering and Electronics, 2018, 40(5): 1109-1117.

基于辅助匹配的1点RANSAC单目视觉导航算法

Monocular visual navigation method with 1-point RANSAC based on aided matching

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 0

编辑推荐

Metrics

本文评价