一种超高分辨率遥感图像融合新算法  

A New Algorithm for Fusing Very High Resolution Remote Sensing Images

作　　者：李旭[1,2] 何明一[1] Michel Roux 卫保国[1] 

机构地区：[1]西北工业大学电子信息学院陕西省信息获取与处理重点实验室,西安710072 [2]国立巴黎高等电信学院信号与图像处理系,巴黎75013

出　　处：《电子与信息学报》2009年第12期2886-2891,共6页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(60572097;60736007);陕西省自然科学基金(2007F22);西北工业大学基础研究计划资助课题

摘　　要：该文针对超高分辨率的全色光图像和多光谱图像的融合,提出了一种基于对应分析的图像融合新算法。该算法在对多光谱数据进行对应分析的基础上,利用冗余小波变换提取出全色光图像的空间细节信息并将其融入到成分空间。实验分别采用IKONOS和QuickBird数据,融合结果的目视效果与客观评价表明,相比现有同类融合方法,该方法能够在提高空间分辨率的同时更好地保持光谱特性,有效地减少了色彩失真的现象。In this paper, a new image fusion approach is presented based on correspondence analysis to sharpen the multispectral images with the panchromatic image. First, the original multispectral images are transformed into component space by correspondence analysis. Then, the panchromatic spatial detail information extracted by the redundant wavelet is injected into the component space. Finally, the fused results are obtained through the inverse transformation. Different IKONOS and QuickBird images have been fused with this new approach. Visual and statistical analyses prove that the proposed method does significantly overcome the color distortion and improve the fusion quality compare to some existing techniques.

关 键 词：图像融合 多光谱图像 全色光图像 对应分析 冗余小波变换 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]