Comparison of Pansharpening Methods Based on Fengyun-4B GHI

Pansharpening aims to improve the spatial resolution of multispectral (MS) images while maintaining their spectral properties, which is very important for military reconnaissance, environmental monitoring, urban planning, and other fields that require high-resolution remote sensing data. With the continuous development of global satellite technology, various new types of sensors have emerged, and the data obtained from different sensors provide valuable resources for scientific research. The fast imager GHI on board the Fengyun-4B star (FY-4B) provides 250m panchromatic (PAN) images and 500m MS images. In order to obtain the desired images at the highest spatial resolution, fusion of the original MS images with the high spatial resolution PAN images using pansharpening techniques to generate high-resolution multispectral images is a feasible way. In this paper, nine commonly used pansharpening algorithms are used to perform synthetic and real experiments on GHI data, and the SFIM method demonstrates excellent performance in both visual and various quantitative metrics evaluations, which are able to accurately retain the spectral properties of the original images while improving the spatial resolution of the images.