首页 | 官方网站   微博 | 高级检索  
     

图结构在航空遥感图像特征点匹配中的应用
引用本文:刘朝霞,邵 峰,景 雨,祁瑞华.图结构在航空遥感图像特征点匹配中的应用[J].计算机工程与应用,2018,54(1):19-24.
作者姓名:刘朝霞  邵 峰  景 雨  祁瑞华
作者单位:1.大连外国语大学 软件学院,辽宁 大连 116044 2.大连海事大学 信息科学技术学院,辽宁 大连 116026
摘    要:为了解决海上航空遥感图像拍摄目标及相机视角变化快,相似特征多导致的匹配不一致问题,并提高匹配算法的效率,提出了一种基于图结构的航空遥感图像特征点匹配算法。该特征匹配过程分为初始匹配和精确匹配两步,首先采用快速特征提取算法SURF进行特征提取并进行初始匹配,然后在精确匹配过程中充分考虑图像的局部空间结构及全局信息,建立K]近邻图结构,用K]近邻结构差异与点集的变换误差作为匹配的收敛条件,以解决图像目标发生变化而导致的K]近邻结构不一致及K]近邻结构相同时仍然存在干扰点两个问题,最终实现快速精确的特征匹配。

关 键 词:特征匹配  SURF特征  结构特征  全局变换误差  图结构  

Application of graph structure in aerial image feature matching
LIU Zhaoxia,SHAO Feng,JING Yu,QI Ruihua.Application of graph structure in aerial image feature matching[J].Computer Engineering and Applications,2018,54(1):19-24.
Authors:LIU Zhaoxia  SHAO Feng  JING Yu  QI Ruihua
Affiliation:1.School of Software, Dalian University of Foreign Languages, Dalian, Liaoning 116044, China 2.Faculty of Information Science & Technology, Dalian Maritime University, Dalian, Liaoning 116026, China
Abstract:The aerial remote sensing image captured on the sea has the characteristics of dynamicity, monotonous and similar patterns, which may result in mismatches in feature matching. In order to remove the mismatches and improve the efficiency, a feature point matching algorithm based on graph structure is proposed. There are two stages in feature matching process, coarse matching and fine matching. In the coarse matching, Speeded Up Robust Features(SURF ) is used to get an initial matching results. In the fine matching, Restricted GTM(RGTM) is proposed. Both local spatial structure and global information are considered. KNN graphs are constructed by the initial points. Then feature matching is converted to graph matching and global transformation error is used to restrict the matching convergence. The experimental results show that the proposed algorithm is more efficient and accurate than GTM.
Keywords:feature matching  Speeded Up Robust Features(SURF)  structure feature  global transformation error  graph structure  
点击此处可从《计算机工程与应用》浏览原始摘要信息
点击此处可从《计算机工程与应用》下载全文
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-23

京公网安备 11010802026262号