PROJECT TITLE :

High Quality Bayesian Pansharpening

ABSTRACT:

Fusion of low-resolution multispectral images with high resolution panchromatic images is known as pansharpening, and the result is a multi-spectral image with high spatial resolution. Using Bayesian theory, a new pansharpening method is proposed in this article. There are three main presuppositions that underlie this algorithm: 1) The geometric information in the pan-sharpened image is identical to that in panchromatic images; 2) The spectral information in both pan-sharpened and panchromatic images is identical; and 3) the neighbouring pixels in each pan-sharpened image channel are similar. The alternating direction approach of multipliers is used to solve our posterior probability model based on the above-mentioned assumptions. The suggested method outperforms the current state-of-the-art pansharpening algorithms in both reduced and full resolution tests. The novel approach is also proven to be effective in preserving information in both spectral and spatial dimensions. It has been demonstrated in additional trials that the proposed method can be extended to hyper-spectral picture fusion with great effectiveness.