An improved cartoon+texture decomposition based pansharpening method

Abstract

Pansharpening is the most widely used fusion method, in the field of remote sensing, to increase spatial information of the multispectral image while preserving spectral signatures. Based on the nature of spatial and spectral information, there is a lack of correlation between them. Therefore, separation of them can be considered as an image decomposition to uncorrelated components. Recently, the cartoon+texture decomposition was used in the pansharpening and decrease spectral distortion. However, details have not been strengthened enough. Therefore, in this paper we aim to use a filter based detail extraction to improve spatial information while mitigate spectral distortion.