首页 | 本学科首页   官方微博 | 高级检索  
     检索      

一种基于图像特征点信息的误匹配点剔除改进算法
引用本文:胡松,李超,王传银,杨明.一种基于图像特征点信息的误匹配点剔除改进算法[J].成都信息工程学院学报,2011,26(5):559-563.
作者姓名:胡松  李超  王传银  杨明
作者单位:1. 成都信息工程学院计算机学院,四川成都,610225
2. 重庆邮电大学计算机学院,重庆,400065
3. 浙江省气象信息网络中心,浙江杭州,310017
基金项目:国家公益性行业(气象)科研专项资助项目(GYHY201106007)
摘    要:误匹配点剔除是基于特征点图像匹配过程的重要一步,关系到图像匹配的精确度。在特征点匹配求精方面,改进基于特征点信息的误匹配点剔除算法,通过自适应特征点邻域半径得到的邻域特征点数量及位置信息进行逐步求精,克服原算法不能处理尺度变换的情形。实验结果表明,该算法简单高效,剔除误匹配点对能力较强,具有尺度、旋转不变性,同时,该算法具有较好的鲁棒性,正确性。能基本满足图像匹配的精度要求。

关 键 词:图像匹配  误匹配点剔除  自适应半径  邻域信息

An Improved Outliers Removal Algorithm Based on Feature Point Information
HU Song,LI Chao,WANG Chuan-yin,YANG Ming.An Improved Outliers Removal Algorithm Based on Feature Point Information[J].Journal of Chengdu University of Information Technology,2011,26(5):559-563.
Authors:HU Song  LI Chao  WANG Chuan-yin  YANG Ming
Institution:1.School of Computer Science and Technology,Chengdu University of Information Technology,Chengdu 610225,China;2.School of Computer Science and Technology,Chongqing University of Post and Telecommunications,Chongqing 400065,China;3.Zhejiang Meteorological Information Network Center,Hangzhou,Zhejiang 310017,China)
Abstract:Outliers removal is an important step in the process of image matching and it influences the accuracy of image matching directly. In terms of refinement process of feature points matching, an improved outliers removal algo-rithm based on feature point is proposed in this paper. The quantity and position information of neighborhood feature points got by self-adaptive radius is used to refine the matching points. And it is invariant to image scaling in the improved algorithm. Experimental results show that the improved algorithm is simple and efficient. The ability to remove the outliers is strong and invariant to image scaling and rotation changes. Furthermore, the good performances of the robustness and correctness can stratify the basic accuracy requirements of image matching.
Keywords:image matching  outliers removal  self-adaptive radius  neighborhood information
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号