SAR图像与多光谱图像融合方法研究  

The Research on the Method of Fusion for SAR and Multi-spectrum Images

作　　者：张红蕾[1] 宋建社[1] 张宪伟[1] 赵延[1] 

机构地区：[1]第二炮兵工程学院信息工程研究所,西安710025

出　　处：《影像技术》2006年第1期48-50,23,共4页Image Technology

基　　金：国家自然科学基金资助项目(批准号:60272022)

摘　　要：本文针对SAR图像分辨率高、纹理信息多、斑噪影响大,而多光谱图像光谱信息丰富、分辨率较小的特点,提出了一种SAR与多光谱图像的数据融合算法。该方法分三步进行:首先使用粗糙集理论选取图像特征,在此基础上,对特征提取后的图像进行配准,配准中采用了数理统计中近似正态分布的方法,配准精度在一个像素以内,最后采用相关性加权的方法对配准后的SA R图像与多光谱图像进行融合。融合实验以LandSat的T M图像与R adsat的SA R图像为例,实验结果表明该方法取得了较好的融合效果。Considering the features of multi-spectrum and SAIL remote sensing images,a new data fusion approach has been proposed in this paper.This method includes three parts：Firstly,the characters of two kinds of images are selected by means of Rough-Set theory.Then,a new image registration method is presented based on the theory of approximately normal distribution.This makes the registration errors of experiment images lower than a pixel.At last,the fusion method based on correhtion weighting approach for SAR and multi-spectrum images is researched.With the images of Landsat TM and Radsat SAIL,experiment results show that the algorithm is available.

关 键 词：合成孔径雷达 图像融合 图像配准 粗糙集 特征提取 

分 类 号：TN911.73[电子电信—通信与信息系统]