Abstract:Pan-sharpening aims to generate high-detail multi-spectral images (HRMS) through the fusion of panchromatic (PAN) and multi-spectral (MS) images. However, existing pan-sharpening methods often suffer from significant performance degradation when dealing with out-of-distribution data, as they assume the training and test datasets are independent and identically distributed. To overcome this challenge, we propose a novel frequency domain-irrelevant feature learning framework that exhibits exceptional generalization capabilities. Our approach involves parallel extraction and processing of domain-irrelevant information from the amplitude and phase components of the input images. Specifically, we design a frequency information separation module to extract the amplitude and phase components of the paired images. The learnable high-pass filter is then employed to eliminate domain-specific information from the amplitude spectrums. After that, we devised two specialized sub-networks (AFL-Net and PFL-Net) to perform targeted learning of the frequency domain-irrelevant information. This allows our method to effectively capture the complementary domain-irrelevant information contained in the amplitude and phase spectra of the images. Finally, the information fusion and restoration module dynamically adjusts the feature channel weights, enabling the network to output high-quality HRMS images. Through this frequency domain-irrelevant feature learning framework, our method balances generalization capability and network performance on the distribution of training dataset. Extensive experiments conducted on various satellite datasets demonstrate the effectiveness of our method for generalized pan-sharpening. Our proposed network outperforms state-of-the-art methods in terms of both quantitative metrics and visual quality, showcasing its superior ability to handle diverse, out-of-distribution data.

Domain-specific knowledge-driven pan-sharpening algorithm

A Scale-Aware Pansharpening Method with Rolling Guidance Filter.

Pan-Sharpening via Multiscale Dynamic Convolutional Neural Network

A New Pansharpening Method Using Multistage Joint Bilateral Filtering

Sparse Representation-Based Detail-Injection Method for Pan-Sharpening

Detail Information Prior Net for Remote Sensing Image Pansharpening

A NOVEL TWO-STAGE GUIDED FILTERING BASED PANSHARPENING METHOD

Frequency Decoupled Domain-Irrelevant Feature Learning for Pan-sharpening

PAN-Guided Band-Aware Multi-Spectral Feature Enhancement for Pan-Sharpening

Detail Injection-Based Deep Convolutional Neural Networks for Pansharpening

Shallow–Deep Convolutional Network and Spectral-Discrimination-Based Detail Injection for Multispectral Imagery Pan-Sharpening

Domain-irrelevant Feature Learning for Generalizable Pan-sharpening

Pansharpening by Combining Enhanced Spatial-Spectral Fidelity and Gradient-Domain Guided Filtering

A Multi-Stage Progressive Pansharpening Network Based on Detail Injection with Redundancy Reduction

Pan-sharpening Via Multi-Scale and Multiple Deep Neural Networks

Pansharpening Based On Details Injection Model And Online Sparse Dictionary Learning

Pyramid Dual Domain Injection Network for Pan-sharpening

Spatial-Frequency Domain Information Integration for Pan-Sharpening

Pan-Sharpening Based On Parallel Pyramid Convolutional Neural Network

Pannet: A Deep Network Architecture For Pan-Sharpening

A Three Stages Detail Injection Network for Remote Sensing Images Pansharpening