基于SIFT与Contourlet变换的高分辨遥感图像配准  

Registration Technique for High-resolution Remote Sensing Images based on SIFT and Contourlet Transform

作　　者：欧阳能钧[1] 李伟彤[1] 韦蔚[2] 潘晴[1] 

机构地区：[1]广东工业大学信息工程学院,广东广州510006 [2]中国地质大学(北京)地球科学与资源学院,北京100083

出　　处：《遥感技术与应用》2013年第1期58-64,共7页Remote Sensing Technology and Application

基　　金：国家自然科学基金项目(61001179);广东省自然科学基金项目(07301038;9451009001002667)

摘　　要：针对高分辨遥感图像特征量较多的情况,提出一种基于SIFT与Contourlet变换相结合的图像配准算法。首先将图像进行Contourlet变换分解成低频和高频子带,对高频子带通过设定合适的阈值来提取图像边缘特征点,对低频子带进行SIFT特征点提取。将两者提取到的特征点分别匹配后得到粗匹配点对,利用随机抽样一致性(RANSAC)选择出精匹配点对,实现图像配准。实验表明:在多源遥感图像配准过程中,与基于非采样Contourlet变换(NSCT)和基于SIFT特征提取相比,该算法能够更准确地提取到特征点,具有更高的运算效率以及匹配率。For high-resolution remote images with tremendous feature, a new method of image registration is proposed by combining contourlet transformation with SIFT. Firstly,the discrete contourlet transformation decomposes the original image into a low-frequency sub-band and several high-frequency sub-band. For high-frequency sub-bands, image edge feature points are extracted by an appropriate threshold. For the low- frequency sub-band, image SIFT feature are extracted by SIFT. Then, coarse control point pairs are searched out with normalized corss-correlation matching and SIFT matching before accurate control point pairs are hunted by RANSAC. The experimental results show that, as for high-resolution remote sensing image registration,this method can extract more acculately control point pairs, get higher computing effi- ciency and matching ratio than these methods based on NSCT or SIFT.

关 键 词：SIFT CONTOURLET变换 遥感图像配准 RANSAC 高分辨率 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]