A NOVEL ADAPTIVE REMOTE SENSING PANSHARPENING ALGORITHM BASED ON THE ICM

Abstract

Abstract. In the paper, a novel Intersecting Cortical Network Model (ICM) based adaptive pansharpening algorithm is proposed to solve the deficiency of spectral distortion and texture detail missing in the remote sensing image fusion. The Shuffled Frog Leaping Algorithm (SFLA) is used in the proposed method to adaptively optimize the ICM model parameters. The fitness function of SFLA is constructed by fusion evaluation index Q4 and SAM, which can generate the irregular optimal segmentation regions. Then, these regions are used to adaptively extract the detail information of the panchromatic image. Finally, the sharpened higher resolution image is obtained with the weighted details and the multispectral upsampling image. Experiments are carried out with the WorldView-2 and GF-2 high-resolution datasets. The experimental results shown that the proposed algorithm performs better compared with the existing pansharpening fusion methods both in the spectral preservation and spatial detail enhancement, which verifies the effectiveness of the algorithm.