基于仿射不变SIFT特征的SAR图像配准  

SAR Image Registration Based on Affine Invariant SIFT Features

作　　者：刘向增[1] 田铮[1,2] 温金环[1] 武建明[1] 张朝阳[1] 

机构地区：[1]西北工业大学应用数学系,西安710129 [2]中国科学院遥感应用研究所,北京100101

出　　处：《光电工程》2010年第11期121-127,共7页Opto-Electronic Engineering

基　　金：国家自然科学基金(60972150;10926197);西北工业大学基础研究基金(JC201053);西北工业大学科技创新基金(2007KJ01033)

摘　　要：针对SAR(Synthetic Aperture Radar)图像全自动配准问题,本文提出一种基于仿射不变SIFT(Scale Invariant Feature Transform)特征的精确配准方法。该方法首先对传统SIFT方法改进构建具有仿射不变性的SIFT描述子,并利用该描述子对提取的控制点进行粗匹配,然后由粗匹配点对的尺度比和方位差及其邻域的灰度相似性构建新的相似矩阵,最后利用SVD(Singular Value Decomposition)方法确定精确匹配点对,求出变换参数从而实现图像的精确配准。实验结果表明该方法优于传统的SIFT方法和SIFT+SVD方法并且可以达到亚像素的配准精度。Referring to the problems of Synthetic Aperture Radar（SAR） image registration,an approach of SAR image registration based on affine invariant Scale Invariant Feature Transform（SIFT） features is proposed.First,the affine invariant SIFT descriptors are constructed by improving the original SIFT descriptors and the control points are rough matched by using the improved SIFT descriptors.Then,a new proximity matrix is built according to the scale ratio and the difference of orientations among the rough matched points and the similarity of the intensity around the rough matched points.Finally,the exact matching points are determined by using Singular Value Decomposition（SVD） method,and the precise image registration is performed.Experimental results show that the proposed algorithm has a better performance than original SIFT algorithm and SIFT＋SVD method,and the accuracy of our algorithm can achieve sub-pixel level.

关 键 词：SIFT SAR图像配准 SVD 相似矩阵 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]