Skip to Main Content
Image fusion techniques can be used to enhance the resolution of multispectral (MS) image, which is helpful for categorisation, recognition as well as other decision making processes. In this paper, a new class of image fusion algorithms is proposed that decomposes images into similar and non-similar information and fuses the corresponding non-similar one. It is based on temporal Fourier analysis. The details of panchromatic (PAN) image are fused with details of MS image by eliminating their similar information which is done by high-pass filtering of their temporal Fourier transform. The cut-off frequency of this filtering is obtained adaptively based on the input MS and PAN images. It results in minimum spectral and spatial distortion. The experimental results demonstrate the superior performance of the proposed method.