基于IHS变换和Curvelet变换的卫星遥感图像融合方法  

A Fusion Method of Satellite Remote Sensing Image Based on IHS Transform and Curvelet Transform

作　　者：肖化超[1] 周诠[1] 郑小松[1] 

机构地区：[1]中国空间技术研究院空间电子信息技术研究院,陕西西安710100

出　　处：《华南理工大学学报（自然科学版）》2016年第1期58-64,共7页Journal of South China University of Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61372175)~~

摘　　要：全色图像和多光谱图像融合是合成高分辨率彩色遥感图像的常用方法之一,高质量的彩色图像需要融合全色图像的高空间分辨率和保持多光谱图像的光谱特性.为提高融合图像的质量,文中提出了一种基于IHS和Curvelet变换的遥感图像融合方法,该方法首先将多光谱图像进行IHS变换,然后将代表空间分辨率的I分量与全色图像进行Curvelet变换,并进行标准差融合,最后逆变换成高质量的彩色图像.实验测试结果表明:文中方法合成图像的信息熵、相对无量纲的全局误差和通用图像质量指标均优于其他方法合成的图像,文中方法融合图像的平均梯度和相关系数与小波变换、PCA变换方法相当;文中方法的融合效果要优于其他方法.The fusion of panchromatic images and multi spectral ones is one of the most commonly-used methods to compose high-resolution color remote sensing images.In order to achieve high-quality color images, it is necessary to fuse the high spatial resolution of panchromatic images and preserve the spectral characteristics of multi spectral images.In this paper, a novel fusion method of remote sensing images on the basis of the IHS transform and the Curvelet transform is proposed to improve the quality of fusion images.In the method, first, the RGB multi spectral images are converted into the IHS images.Then, the panchromatic image and the I component refering to spatial resolution are handled through the Curvelet transform, and are fused by means of the standard difference method. Finally, the high-quality color images are achieved through the inverse transforms.Experimnetal results show that the images fused through the proposed method are superior to those through the other methods in terms of entropy, relatively-dimensionless global error and universal image quality index, and are equivalent to those through the wavelet transform and the PCA transform in terms of average gradient and correlation coefficient.Overall, the pro-posed method has better fusion performance than the other three typical methods.

关 键 词：图像融合 多光谱图像 全色图像 CURVELET变换 IHS变换 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]