Remote Sensing Image Dehazing: A Systematic Review of Progress, Challenges, and Prospects

For the purposes of this review, we adopt an inclusive definition of “dehazing” that encompasses all methodologies designed to mitigate the effects of fog, haze, and optically cloud layers. In this review, we have systematically examined over 200 papers :scroll::scroll::scroll:, summarizing and analyzing more than 100 Remote Sensing Image Dehazing methods.

:heart_eyes_cat: If this work is helpful for you, please help star this repo. Thanks!

:mega: News

• 2026/7/6: Added 1 ECCV 2026 paper
• 2026/3/11: Added 4 TGRS 2026 papers, 2 JSTARS 2026 papers
• 2026/3/9: Our paper has been officially published in the ISPRS Journal of Photogrammetry and Remote Sensing. The preprint version is available on arXiv. :balloon::balloon::balloon:
• 2026/1/30: Added 1 TIP 2025 paper, 1 TCSVT 2025 paper, 1 GRSL 2025 paper, 2 TGRS 2025 papers
• 2025/12/15: Added 1 TGRS 2025 paper, 1 JSTARS 2025 paper
• 2025/12/05: :trophy: More than 100 methods have been included !
• 2025/12/05: Added 1 ACM MM 2025 paper, 1 ICASSP 2025 paper, 2 JSTARS 2025 papers, 2 TGRS 2025 papers, 1 Remote Sensing 2020 paper, 1 IJCNN 2025 paper
• 2025/11/09: Added 2 datasets: HyperDehazing, RRSHID.
• 2025/10/25: Added a Reproducibility Checklist.
• 2025/08/29: Added 2 TGRS 2025 papers, 1 TGRS 2024 paper.
• 2025/07/26: Added 2 JSTARS 2025 papers, 1 GRSL 2025 paper.
• 2025/07/22: Added 3 TGRS 2025 papers, 1 EAAI 2025 paper, 1 ISPRS P&RS 2024 paper and 1 Signal Processing 2025 paper.
• 2025/06/28: Paper submitted.
• 2025/05/15: Added 2 CVPR 2025 papers.

:pizza: Introduction

Remote sensing images (RSIs) are frequently degraded by haze, fog, and thin clouds, which obscure surface reflectance and hinder downstream applications. This study presents the first systematic and unified survey of RSIs dehazing, integrating methodological evolution, benchmark assessment, and physical consistency analysis. We categorize existing approaches into a three-stage progression: from handcrafted physical priors, to data-driven deep restoration, and finally to hybrid physical-intelligent generation, and summarize more than 30 representative methods across CNNs, GANs, Transformers, and diffusion models. To provide a reliable empirical reference, we conduct large-scale quantitative experiments on five public datasets using 12 metrics, including PSNR, SSIM, CIEDE, LPIPS, FID, SAM, ERGAS, UIQI, QNR, NIQE, and HIST. Cross-domain comparison reveals that recent Transformer and diffusion-based models improve SSIM by 12%–18% and reduce perceptual errors by 20%–35% on average, while hybrid physics-guided designs achieve higher radiometric stability. A dedicated physical radiometric consistency experiment further demonstrates that models with explicit transmission or air light constraints reduce color bias by up to 27%. Based on these findings, we summarize open challenges: dynamic atmospheric modeling, multimodal fusion, lightweight deployment, data scarcity, and joint degradation, and outline promising research directions for future development of trustworthy, controllable, and efficient (TCE) dehazing systems. In addition, we discuss key technical challenges in Fig.2, such as dynamic atmospheric modeling, multi-modal data fusion, lightweight model design, data scarcity, and joint degradation scenarios, and propose future research directions.

Fig 1. Taxonomy of Remote Sensing Image Dehazing Methods.

Content:

0. Remote Sensing Image Datasets
1. Traditional Remote Sensing Image Restoration Methods
2. Deep Convolution for Remote Sensing Image Dehazing
3. Adversarial Generation for Remote Sensing Image Dehazing
4. Vision Transformer for Remote Sensing Image Dehazing
5. Diffusion Generation for Remote Sensing Image Dehazing
6. Current Challenges and Future Prospects
7. Evaluation

:open_file_folder: Remote Sensing Image Datasets:

1. Traditional Image Enhancement and Physics Model for Remote Sensing Image Dehazing:

:rocket::rocket::rocket:Update (in 2026-3-11) :balloon:

2. Deep Convolution for Remote Sensing Image Dehazing:

:rocket::rocket::rocket:Update (in 2026-3-11) :balloon:

3. Adversarial Generation for Remote Sensing Image Dehazing:

:rocket::rocket::rocket:Update (in 2026-3-11) :balloon:

4. Vision Transformer for Remote Sensing Image Dehazing:

:rocket::rocket::rocket:Update (in 2026-3-11) :balloon:

5. Diffusion Generation for Remote Sensing Image Dehazing:

:rocket::rocket::rocket:Update (in 2026-7-6) :balloon:

6. :surfer: Current Challenges and Future Prospects

Fig 2. Future prospects for RSI dehazing: Trustworthy, controllable, and efficient (TCE) remote sensing dehazing systems.

:bar_chart: Evaluation:

For evaluation on Dehazed results, modify 'test_original' and 'test_restored' to the corresponding path

python evaluate.py --train_folder [restored image path] --target_folder [ground-truth image path]

Make sure the file structure is consistent with the following:

dataset
├── Restored
│   ├── RICE
│   ├── RRSHID-M
│   ├── RRSHID-TK
│   ├── RRSHID-TN
│   ├── SH-M
│   ├── SH-TK
│   └── SH-TN
│       └── 1.png, 2.png, ...
|
├── Ground-truth
│   ├── RICE-GT
│   ├── RRSHID-M-GT
│   ├── RRSHID-TK-GT
│   ├── RRSHID-TN-GT
│   ├── SH-M-GT
│   ├── SH-TK-GT
│   └── SH-TN-GT
│       └── 1.png, 2.png, ...

Table 1. Quantitative performance at PSNR (dB) and SSIM of remote sensing image restoration algorithms evaluated on the SateHaze1k (SH-TN, SH-M, SH-TK) and RICE datasets.

Go to Reproducibility Checklist

:books: Citation:

• If you find our survey paper (ISPRS, arXiv) and evaluation code are useful, please cite the following paper:

@article{zhou2026remote,
  title={Remote Sensing Image Dehazing: A Systematic Review of Progress, Challenges, and Prospects},
  author={Heng Zhou and Xiaoxiong Liu and Zhenxi Zhang and Jieheng Yun and Chengyang Li and Yunchu Yang and Dongyi Xia and Chunna Tian and Xiao-Jun Wu},
  journal={ISPRS Journal of Photogrammetry and Remote Sensing},
  volume = {235},
  pages = {318-348},
  year = {2026},
  publisher={Elsevier}
}

%

:clap::clap::clap: Thanks to the above authors for their excellent work！