Real-World Remote Sensing Image Dehazing: Benchmark and Baseline

2025-03-23

AI Summary

To address inaccurate atmospheric modeling and severe color distortion in real-world remote sensing image dehazing, as well as the poor generalizability of existing methods trained predominantly on synthetic data, this paper introduces RRSHID—the first large-scale, real-world paired remote sensing dehazing dataset—and proposes MCAF-Net, a model specifically designed for real-scene dehazing. MCAF-Net incorporates three novel components: a Multi-branch Feature Interaction Block (MFIBA) for physically grounded atmospheric estimation, a self-supervised Color-Spectrum Attention Module (CSAM) for chromatic fidelity preservation, and a Multi-scale Adaptive Feature Fusion Module (MFAFM) for structural detail recovery. Jointly, these modules ensure physical plausibility, accurate color reproduction, and robust structural restoration. Extensive experiments demonstrate state-of-the-art performance on both real and synthetic benchmarks, with significant improvements in dehazing quality and cross-domain generalization. The code and RRSHID dataset are publicly released.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Machine Learning: Multi-instance/Multi-view Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Remote Sensing Image Dehazing (RSID) poses significant challenges in real-world scenarios due to the complex atmospheric conditions and severe color distortions that degrade image quality. The scarcity of real-world remote sensing hazy image pairs has compelled existing methods to rely primarily on synthetic datasets. However, these methods struggle with real-world applications due to the inherent domain gap between synthetic and real data. To address this, we introduce Real-World Remote Sensing Hazy Image Dataset (RRSHID), the first large-scale dataset featuring real-world hazy and dehazed image pairs across diverse atmospheric conditions. Based on this, we propose MCAF-Net, a novel framework tailored for real-world RSID. Its effectiveness arises from three innovative components: Multi-branch Feature Integration Block Aggregator (MFIBA), which enables robust feature extraction through cascaded integration blocks and parallel multi-branch processing; Color-Calibrated Self-Supervised Attention Module (CSAM), which mitigates complex color distortions via self-supervised learning and attention-guided refinement; and Multi-Scale Feature Adaptive Fusion Module (MFAFM), which integrates features effectively while preserving local details and global context. Extensive experiments validate that MCAF-Net demonstrates state-of-the-art performance in real-world RSID, while maintaining competitive performance on synthetic datasets. The introduction of RRSHID and MCAF-Net sets new benchmarks for real-world RSID research, advancing practical solutions for this complex task. The code and dataset are publicly available at url{https://github.com/lwCVer/RRSHID}.

Problem

• Addressing real-world remote sensing image dehazing challenges
• Bridging domain gap between synthetic and real hazy images
• Mitigating color distortions in degraded remote sensing images

Innovation

• Large-scale real-world hazy image dataset RRSHID
• Multi-branch feature integration MFIBA block
• Color-calibrated self-supervised attention CSAM