A Multispectral and Panchromatic Image Fusion Method Based on NSST and PCA

Abstract

In order to improve spectral distortion and the problem of lacking details, a pixel-level image fusion method based on the Non-subsampled Shearlet Transform (NSST) is proposed. Firstly, the Principal Component Analysis (PCA) transform is applied to the multispectral image with low spatial resolution to obtain the relatively independent component containing the most abundant information, which can represent the original multispectral images. Secondly, the panchromatic image after linear stretching and the first PCA transform component of multispectral image are respectively decomposed into different coefficients to better approach the image edge according to NSST, which is translation invariant. Thirdly, fusion rules for different decomposed coefficients are established to merge the source images. Finally, the fusion image is obtained by inverse NSST and inverse PCA. The experimental results show the efficacy of the proposed method in details preservation and the spectral distortion reduction.