改进的IHS变换和小波变换相结合的图像融合

在分析IHS变换时发现，IHS变换中亮度分量的计算是从红绿蓝三种颜色中平均提取三分之一作为总亮度，而人眼对绿色最敏感（占60％）、其次是红色（占30％）、再者是蓝色（占10％），故本文对亮度分量的计算公式做了修改。在此基础上，提出了一种改进的IHS变换和小波变换相结合的图像融合方法。首先对多光谱图像进行改进的IHS变换，然后利用小波变换分别对1分量和全色高分辨率图像做小波分解，并对小波低频系数基于局部能量进行融合，小波高频系数基于局部方差进行融合，最后进行IHS逆变换，得到融合结果图像。结果表明，本文提出的方法在保持光谱源图像特性方面有优势，也较多地保留了源图像的空间细节信息，融合图像的扭曲程度小。

Image Fusion in Combination of the Improved IHS Transform and Wavelet Transform

In the analysis and IHS transform we found that total intensity is obtained through a third from each red, green and blue in calculation of intensity of IHS transformation. As the human eye is the most sensitive to green （60%）, followed by red （30%）, moreover is blue （10%）, therefore, the formula of intensity component has been modified in this paper. On this basis, a image fusion method of combined the improved IHS transform with wavelet transform has been presented. First, the improved IHS transform for multispectral image is conducted, and the I component and high-resolution panchromatic image are decomposed using wavelet transform, respectively, and the low frequency and high frequency coefficient of high-resolution panchromatic image are fused with those of the I component based on local energy and on local variance, respectively. Finally, IHS inverse transformation is conducted, and the fusion result images are obtained. Results show that the proposed method in this paper has an advantage in keeping the source image spectral characteristics and also keep the spatial fine structure of the source image, and decrease distortion degree of fused image.