基于非亚采样Contourlet和SWT的多光谱图像和全色图像的融合算法

Fusion of Multispectral and Panchromatic Satellite Images Based on Nonsubsampled Contourlet Transform and Stationary Wavelet Transform

该文研究了多尺度几何分析工具非亚采样Contourlet变换(NSCT),提出一种新的全色图像和多光谱图像融合的方法。该方法首先对全色图像和进行过IHS变换的多光谱图像的亮度分量进行NSCT变换,对于二者的低频近似系数再进行平稳小波变换(SWT)并融合,进一步提高融合图像的空间信息量,对于高频细节系数,采用基于局部平均梯度的方法进行融合,经过逆NSCT得到融合图像。实验结果表明,该文提出的方法在保留多光谱图像的光谱信息的同时,增强了融合图像的空间细节表现能力,提高了信息量,并且优于传统的基于IHS变换、小波变换、双树复小波变换及Contourlet变换的融合方法,该方法是有效可行的。

This paper researches the multiscal geometry analysis tool, NonSubsampled Contourlet Transform （NSCT）, and proposes a new method of fusing panchromatic and multispectral images based on NSCT and Stationary Wavelet Transform （SWT）. This method first performs NSCT on panchromatic image and the intensity component of multispectral image with IHS, then fuses the low frequency coefficients with SWT for further spatial information extraction, and fuses the high frequency coefficients based on local average gradient fusion rule, finally a fused image is obtained through inversing NSCT. The experimental results show that the proposed method can effectively preserve spectral information and improve spatial information of the fused image, and outperforms the traditional IHS, wavelet, DT-CWT and contourlet methods.