Performance Analysis of Pansharpening Algorithms Based on Guided Filtering Applied to KOMPSAT-3/3A Satellite Imagery in Coastal Areas

Abstract

Choi, J.; Park, H.; Seo, D., and Han, Y., 2018. Performance analysis of pansharpening algorithms based on guided filtering applied to KOMPSAT-3/3A satellite imagery in coastal areas. In: Lee, J.L.; Yoon, J.-S.; Cho, W.C.; Muin, M., and Lee, J. (eds.), The 3rd International Water Safety Symposium. Journal of Coastal Research, Special Issue No. 91, pp. 431-435. Coconut Creek (Florida), ISSN 0749-0208.Image fusion algorithms, also known as pansharpening algorithms, can improve the spatial resolution of multispectral images using high-resolution original panchromatic images. Pansharpening algorithms are used in preprocessing steps for image classification, change detection, and image interpretation, and they have been used in various marine applications. In this study, we applied various guided filtering (GF)-based pansharpening techniques to multispectral images with high spatial resolutions in coastal areas and evaluated the quality of the resulting pansharpened images. Specifically, we analyzed the results of GF-based pansharpening techniques according to the characteristics of the guidance and filtering input images used in GF. The experimental results show that the spatial and spectral qualities of GF-based pansharpened images in the studied coastal areas are determined by the characteristics of the filtering input image. Additionally, the results of GF-based pansharpening display spectral distortion and spatial clarity characteristics similar to those of component substitution (CS) and multiresolution analysis (MRA)-based techniques. The obtained GF-based pansharpening results can be used in various coastal applications.