结合Contourlet和HSI变换的组合优化遥感图像融合方法  

A Combinatorial Optimization Method for Remote Sensing Image Fusion with Contourlet and HSI Transform

作　　者：宋梦馨[1] 郭平[1] 

机构地区：[1]北京师范大学图形图像与模式识别实验室,北京100875

出　　处：《计算机辅助设计与图形学学报》2012年第1期83-88,共6页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(90820010;60911130513)

摘　　要：针对遥感图像中全色图像与多光谱图像融合问题,提出一种组合优化图像融合方法——COFM.通过HSI变换获取多光谱图像的亮度分量后,采用Contourlet变换对全色图像和多光谱图像的亮度分量进行分解,分别获取其高频和低频子图;对高频子图提取分形特征,采用取最大的融合规则进行融合;对低频子图提取能量特征后采用第二代非支配排序遗传算法选择融合权值;然后使用加权模型对其进行融合.实验结果表明,COFM的融合效果优于传统图像融合方法,能够在提升图像空间分辨率的同时较好地保留光谱信息.To deal with the problem of panchromatic and multispectral remote sensing images fusion, a combinatorial optimization image fusion method （COFM） is proposed in this paper. First, the HSI transform is used to obtain the intensity component of a multispectral image; Second, Contourlet transform is adopted to decompose the panchromatic image and the intensity component of the multispectral image, respectively. Consequently the high- and low-frequency of sub-images can be obtained. For high-frequency sub-images, fractal features are extracted and the maximum fusion rule is utilized to fuse these features. For low-frequency sub-images, energy features are extracted, and the non-dominated sorting genetic algorithm-Ⅱ （NSGA-Ⅱ） is applied to select the weight coefficients before sub-images are fused with weighted model. The experimental results demonstrate that the performance of COFM is better than that of traditional approaches. COFM can increase the spatial resolution of the fused image while preserving original image＇s spectral information.

关 键 词：遥感图像融合 组合优化 分形特征 第二代非支配排序遗传算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]