Remote sensing image fusion using transform domain based on optimization algorithms

Abstract

With the rapid development of technology today in different fields like military, medicine, robotics, remote sensing, finding underground sources, target tracking, target identification, microscopic imaging, and security applications are need clearer and more meaningful images. Sometimes images from a single sensor are not enough for analysis. For this reason, images taken by different sensors and different features are used together. For this purpose, image fusion (Pan-sharpening) has got great importance. The image obtained by fusion methods contains more meaningful and clear information. In this study, metaheuristic algorithms are used to sharpen the Multispectral (MS) image with a Panchromatic (Pan) image in remote sensing. In this study, the coefficients obtained from Curvelet and Laplacian Pyramid transformations are using with weights that are generated by Particle Swarm Optimization (PSO) and Bat Algorithm (BAT) in the fusion process. A fused image has been successfully achieved by preserving the spatial information of the high-resolution Pan image and the color information of the low-resolution MS image. The obtained results have had a clearer, brighter, and richer edge information. Visual and quantitative comparison of the obtained results were also evaluated.