Summary
Remote sensing images often look blurry or discolored because of haze in the atmosphere, which makes it hard to see important details. The researchers designed a special computer network called DPSF-Net that uses both regular pictures and extra haze-related maps to better understand and remove haze. This network processes image information in both normal and frequency (Fourier) forms, helping it separate large haze areas from real objects. Their method improves image clarity and color accuracy while being efficient to run, showing promising results on real and synthetic images. This approach could help in clearer satellite image analysis, which is important for things like environmental monitoring.
remote sensingimage dehazingdark channel priorspatial-domain featuresfrequency-domain featuresFourier transformattention modulemulti-scale fusiondeep learninghaze removal
Authors
Mei Lu, Shangliang Shao, Shanliang Yao
Abstract
Real-world remote sensing image dehazing (RSID) remains challenging because atmospheric scattering, spatially non-uniform haze and colour distortion jointly degrade structural and spectral information. Most deep learning methods rely on RGB inputs and spatial-domain feature extraction, which limits their ability to separate global background haze from local surface details. Here, we propose DPSF-Net, a dual-prior spatial-frequency network built on MCAF-Net for real-world RSID. The network uses hazy RGB images and dark channel prior (DCP) maps as joint inputs, allowing physical degradation cues to guide end-to-end feature learning. A spatial-frequency residual interaction block introduces a FourierUnit branch into multi-directional spatial interaction to model large-scale haze components. A prior-guided feature attention module adaptively fuses prior and attention features to reduce colour shift and structural distortion. A selective kernel complementary fusion module screens multi-scale skip features through bidirectional residual complementary gating and selective kernel fusion. Extensive experiments demonstrate that DPSF-Net achieves state-of-the-art performance on the real-world RRSHID remote sensing image dehazing benchmark and remains competitive across multiple synthetic datasets. Moreover, the proposed method strikes a favourable balance among restoration quality, parameter count and computational complexity, supporting the effectiveness of dual-prior spatial-frequency modelling.