Remote Sensing Image Fusion via Hybrid Image Decomposition with Spatial Frequency Motivated PCNN

Pan sharpening is pivotal for getting a composite image which consists of the information related to spatial and spectral. In this paper, pan-sharpening method based on co-occurrence filtering with spatial frequency motivated PCNN is proposed. The three parts of panchromatic (PAN) image, i.e., small- and large-scale images, a base image have been obtained through hybrid of co-occurrence and Gaussian filtering (CoF-GF) decomposition. Next, intensity, saturation and hue components of multispectral (MS) image have been obtained through HSI transform. Thirdly, PCNN modulated with spatial frequency has been used to merge the base images and MS image intensity component. Finally, fused output has been reconstructed by an inverse HSI applied on addition of small-scale, large-scale and fused base image. Experiments in three datasets have been validated that the proposed outperforms most of the recently suggested methods.