一种用于组合导航的SAR图像快速匹配算法  

A fast SAR image matching method for integrated navigation

作　　者：刘瑞华[1] 王彦光[1] 

机构地区：[1]中国民航大学电子信息工程学院,天津300300

出　　处：《光电子．激光》2010年第5期762-766,共5页Journal of Optoelectronics·Laser

基　　金：国家"863"计划资助项目(2006AA12Z313)

摘　　要：提出了基于合成孔径雷达(SAR)图像匹配技术的惯性导航系统(INS)/SAR组合导航方法。首先将计算机视觉领域的快速鲁棒性特征(SURF)算子用于对SAR图像的局部特征描述,然后对兴趣点建立快速索引,并用阈值为0.6的最近邻法则进行快速的初匹配,最后使用随机抽样一致(RSC)进一步去除错误的匹配点对。实验中,用2组真实的SAR数据对算法进行仿真并与尺度不变特征变换(SIFT)算子进行比较,结果显示,两者的鲁棒性相差很小,但在匹配时间上本文算法明显优于SIFT算子,本文算法更适用于组合导航的SAR图像匹配。Based on the fast SAR image matching method,a method combining integrated navigation system（INS） and SAR is proposed for improving the integrated navigation system′s capability.First,the computer version′s SURF （speed up robust feature） method is used to describe the local features of SAR images.Then rapid-index of the interested points is established,and the nearest neighborhood method with the threshold of 0.6 is used during the process of initial matching.Finally,the RSC is used to remove the virtual matching point pairs.Two sets of real SAR data are used in the experiments.The results show that though the robustness difference of the two methods is tiny,the proposed method is greatly better than the SIFT method in time consuming.So we can draw a conclusion that the algorithm proposed in this paper is more suitable for integrated navigation in SAR image matching.

关 键 词：惯性导航系统(INS) 合成孔径雷达(SAR) 图像匹配 快速鲁棒特征(SURF) 快速索引 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]