Towards UAV Image Dehazing: A UAV Atmospheric Scattering Model, Benchmark, and Geometry-Aware Deep Unfolding Network

2026-06-15

AI Summary

This work addresses the challenge of severe spatially non-uniform haze in drone-captured images—particularly dense accumulation in the upper regions—which significantly obscures distant scene details. Conventional atmospheric scattering models are ill-suited for such scenarios, and paired real-world hazy/clean data are scarce. To overcome these limitations, the authors propose a UAV-specific atmospheric scattering model (UASM) and introduce a geometry-aware deep unfolding network, GP-DUN, which uniquely integrates flight altitude, pitch angle, and other geometric parameters into a physically grounded dehazing framework to ensure geometric consistency. The method synergistically combines the UASM, geometry-aware gradient descent, a latent geometry estimator, and a Pooling-Expert proximal mapping module. Evaluated on the newly curated UASM-HazeSet—the first real-world benchmark for UAV dehazing—and additional authentic datasets, the approach substantially outperforms existing methods in recovering structural integrity and textural details.

Abstract

In UAV applications, haze significantly obscures distant details and weaken structural information, hindering the recovery of details. Current UAV scenarios still face two key challenges: (i) paired hazy/clean images from the real world are unobtainable, while the classical atmospheric scattering model is inadequate for modeling the spatially non-uniform haze in UAV imagery; (ii) existing dehazing methods struggle to remove the heavy haze accumulated in the upper regions of UAV images. To address these issues, we first propose a UAV Atmospheric Scattering Model (UASM), which explicitly incorporates flight altitude, viewing pitch, and extinction to characterize the non-uniform haze distribution in UAV imaging. Based on UASM, we develop a physics-driven dehazing framework, termed Geometry-aware Proximal Deep Unfolding Network (GP-DUN). Specifically, GP-DUN consists of three key modules: a Latent Geometry Estimator (LGE) that infers transmittance consistent with UAV imaging geometry, a Geometry-aware Gradient Descent Module (GeoGDM) that embeds UASM into the data-fidelity term and performs physics-consistent closed-form updates, and an Pooling-Expert Proximal Mapping Module (PE-PMM) that learns an implicit prior to restore textures and structures beyond the capability of explicit physical modeling. In addition, we further construct UASM-HazeSet, which provides controllable paired synthetic data together with 2,285 real UAV haze images for testing. Extensive experiments show that GP-DUN consistently outperforms existing methods on both UASM-HazeSet and real UAV haze benchmarks.

Problem

• UAV image dehazing
• non-uniform haze
• atmospheric scattering model
• real-world paired data
• heavy haze removal

Innovation

• UAV dehazing
• atmospheric scattering model
• geometry-aware deep unfolding
• non-uniform haze
• physics-driven network