Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (28): 155-157.DOI: 10.3778/j.issn.1002-8331.2008.28.052

• 图形、图像、模式识别 • Previous Articles     Next Articles

Feature point matching based on bidirectional maximal correlation and parallactic restriction

XU Wei,ZHANG Mao-jun,XIONG Zhi-hui,WANG Wei   

  1. College of Information System and Management,National University of Defense Technology,Changsha 410073,China
  • Received:2008-03-26 Revised:2008-06-19 Online:2008-10-01 Published:2008-10-01
  • Contact: XU Wei


徐 玮,张茂军,熊志辉,王 炜   

  1. 国防科技大学 信息系统与管理学院,长沙 410073
  • 通讯作者: 徐 玮

Abstract: This paper proposed a matching method of feature point based on bidirectional maximal correlation and parallactic restriction,which firstly improves the precision of feature point matching by utilizing the strategy of bidirectional matching,and then eliminates false matching point based on parallactic restriction.Experimental results show that this matching method is not only easy and efficient,but also precise for matching.

Key words: feature point, feature point matching, parallactic restriction

摘要: 提出一种基于双向最大相关与视差约束的特征点匹配方法,首先利用双向匹配策略来提高匹配精度,然后基于视差约束条件进一步剔除误匹配点,实验证明该算法不仅简单高效而且匹配精度高。

关键词: 特征点, 特征点匹配, 视差约束