基于区域特征的改进IHS图像融合算法

摘要

An image fusion algorithm based on regional features and IHS wavelet transform is proposed.Firstly,having IHS transform for the multispectral image and having histogram matching for the high resolution image,then introducing regional features criterion when the different frequency components are treated after wavelet transform.Low-frequency part using the weighted averaging operator method,high-frequency part using the improved weighted regional variance method.Subjective judgment and objective analysis both show that the image quality of this paper is better than the one of IHS transform or wavelet transform.%提出了一种基于区域特征的结合IHS变换和小波变换的图像融合算法,首先分别对多光谱图像和高分辨率全色图像进行IHS变换和直方图匹配,对I分量和调整后的高分辨率图像进行小波分解,然后对不同频率分量进行融合时引入区域特征准则,低频部分采用加权平均算子融合法,高频部分采用改进的区域方差加权融合法,最后结合主观判断和客观定量分析验证该算法,实验结果表明该算法的融合图像的质量要优于传统的IHS变换法和小波变换法。