Pan-sharpening via High-pass Modification Convolutional Neural Network

TL;DR

This paper introduces a novel deep learning approach for pan-sharpening that enhances spatial details while reducing spectral distortion by learning high-pass information through a specialized convolutional neural network block.

Contribution

The paper proposes a high-pass modification block within a CNN for pan-sharpening, along with a perceptual loss function optimized in the near-infrared space, improving spatial and spectral quality.

Findings

Outperforms state-of-the-art methods quantitatively

Produces visually appealing pan-sharpened images

Open-source implementation available

Abstract

Most existing deep learning-based pan-sharpening methods have several widely recognized issues, such as spectral distortion and insufficient spatial texture enhancement, we propose a novel pan-sharpening convolutional neural network based on a high-pass modification block. Different from existing methods, the proposed block is designed to learn the high-pass information, leading to enhance spatial information in each band of the multi-spectral-resolution images. To facilitate the generation of visually appealing pan-sharpened images, we propose a perceptual loss function and further optimize the model based on high-level features in the near-infrared space. Experiments demonstrate the superior performance of the proposed method compared to the state-of-the-art pan-sharpening methods, both quantitatively and qualitatively. The proposed model is open-sourced at https://github.com/jiaming-wang/HMB.