多尺度非对称注意力遥感去雾Transformer

doi:10.16088/j.issn.1001-6600.2025061001

摘要/Abstract

摘要： 雾霾干扰会导致遥感图像结构模糊、细节丢失,严重影响下游视觉任务的准确性。为此,本文提出一种异构增强的遥感图像去雾网络,从空间结构建模与频率信息整合2个层面提升特征恢复能力。具体而言,设计多尺度非对称注意力Transformer模块,引入方向感知机制以增强模糊边缘与纹理细节的建模;同时构建基于小波变换高低频自适应增强模块,使用Haar小波分解分离频域信息,分别通过高频与低频子模块强化边缘轮廓与结构表达。2个模块分别嵌入特征提取与融合阶段,协同缓解传统方法方向性建模不足与高频特征易丢失等问题。在保持低计算开销的前提下,本文方法在HAZE1K与RICE数据集上的平均PSNR/SSIM性能分别达到24.993 6/0.909 9与33.180 2/0.894 2,在细节恢复方面表现出显著优势。

关键词: 遥感图像去雾, Transformer, 非对称注意力, 高低频特征增强, 小波变换, 方向感知建模, 深度学习

Abstract: Haze interference can cause blurred structures and loss of details in remote sensing images, severely compromising the accuracy of downstream visual tasks. To address this challenge, this paper proposes a heterogeneous enhancement remote sensing image dehazing network that improves feature restoration from two perspectives: spatial structure modeling and frequency information integration. Specifically, a multi-scale asymmetric attention transformer module is designed, incorporating a direction-aware mechanism to enhance the modeling of blurred edges and texture details. In parallel, a wavelet-based adaptive high-low frequency enhancement module is constructed, utilizing Haar wavelet decomposition to separate frequency-domain information, where high-frequency and low-frequency submodules are employed to reinforce edge contours and structural representations, respectively. These two modules are embedded in the feature extraction and feature fusion stages, collaboratively addressing the limitations of traditional methods in directional modeling and high-frequency feature preservation. With low computational overhead, the proposed method achieves average PSNR/SSIM scores of 24.993 6/0.909 9 on the HAZE1K dataset and 33.180 2/0.894 2 on the RICE dataset, demonstrating significant advantages in detail restoration.

Key words: remote sensing image dehazing, transformer, asymmetric attention, high-low frequency feature enhancement, wavelet transform, direction-aware modeling, deep learning

中图分类号: TP391.41;TP751

王旭阳, 梁宇航. 多尺度非对称注意力遥感去雾Transformer[J]. 广西师范大学学报（自然科学版）, 2026, 44(2): 77-89.

WANG Xuyang, LIANG Yuhang. Multi-scale Asymmetric Attention Transformer for Remote Sensing Image Dehazing[J]. Journal of Guangxi Normal University(Natural Science Edition), 2026, 44(2): 77-89.

参考文献

[1] DALLA MURA M, BENEDIKTSSON J A, WASKE B, et al.Morphological attribute profiles for the analysis of very high resolution images[J].IEEE Transactions on Geoscience and Remote Sensing, 2010, 48(10):3747-3762.DOI:10.1109/TGRS.2010.2048116.
[2] BRUNNER D, LEMOINE G, BRUZZONE L.Earthquake damage assessment of buildings using VHR optical and SAR imagery[J].IEEE Transactions on Geoscience and Remote Sensing, 2010, 48(5):2403-2420.DOI:10.1109/TGRS.2009.2038274.
[3] BERMAN D, TREIBITZ T, AVIDANS.Non-local image dehazing[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).Los Alamitos, CA:IEEE Computer Society, 2016:1674-1682.DOI:10.1109/CVPR.2016.185.
[4] BERMAN D, TREIBITZ T, AVIDANS.Air-light estimation using haze-lines[C]//2017 IEEE International Conference on Computational Photography (ICCP).Piscataway, NJ:IEEE, 2017:115-123.DOI:10.1109/ICCPHOT.2017.7951489.
[5] HE K M, SUN J, TANG X O.Single image haze removal using dark channel prior[J].IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(12):2341-2353.DOI:10.1109/TPAMI.2010.168.
[6] ZHU Q S, MAI J M, SHAO L.A fast single image haze removal algorithm using color attenuation prior[J].IEEE Transactions on Image Processing, 2015, 24(11):3522-3533.DOI:10.1109/TIP.2015.2446191.
[7] LEVIN A, WEISS Y, DURAND F, et al.Understanding and evaluating blind deconvolution algorithms[C]//2009 IEEE Conference on Computer Vision and Pattern Recognition.Piscataway, NJ:IEEE, 2009:964-1971.DOI:10.1109/CVPR.2009.5206815.
[8] SHAN Q, JIA J Y, AGARWALA A.High-quality motion deblurring from a single image[J].ACM Transactions on Graphics, 2008, 27(3):1-10.DOI:10.1145/1360612.1360672.
[9] 陈云龙, 姜钦霞, 邓飞凤, 等.基于多尺度Retinex算法的电力巡检可见光遥感图像去雾自动化增强[J].自动化与仪表, 2024, 39(12):88-91, 95.DOI:10.19557/j.cnki.1001-9944.2024.12.019.
[10] CHEN Z X, HE Z W, LU Z M.DEA-Net:single image dehazing based on detail-enhanced convolution and content-guided attention[J].IEEE Transactions on Image Processing, 2024, 33:1002-1015.DOI:10.1109/TIP.2024.3354108.
[11] ENGIN D, GENC A, EKENEL H K.Cycle-Dehaze:enhanced CycleGAN for single image dehazing[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).Los Alamitos, CA:IEEE Computer Society, 2018:938-946.DOI:10.1109/CVPRW.2018.00127.
[12] TAO X, GAO H Y, SHEN X Y, et al.Scale-recurrent network for deep image deblurring[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Los Alamitos, CA:IEEE Computer Society, 2018:8174-8182.DOI:10.1109/CVPR.2018.00853.
[13] QIN X, WANG Z L, BAI Y C, et al.FFA-Net:feature fusion attention network for single image dehazing[J].Proceedings of the AAAI Conference on Artificial Intelligence, 2020, 34(7):11908-11915.DOI:10.1609/aaai.v34i07.6865.
[14] RAJASEKARAN G, ABITHA V, VAISHNAVI S M.Image dehazing algorithm based on artificial multi-exposure image fusion[J].Multimedia Tools and Applications, 2023, 82(26):41241-41251.DOI:10.1007/s11042-023-15210-6.
[15] LU L P, XIONG Q, XU B R, et al.MixDehazeNet:mix structure block for image dehazing network[C]//2024 International Joint Conference on Neural Networks (IJCNN).Piscataway, NJ:IEEE, 2024:1-10.DOI:10.1109/IJCNN60899.2024.10651326.
[16] 孙航, 方帅领, 但志平, 等.层级特征交互与增强感受野双分支遥感图像去雾网络[J].遥感学报, 2023, 27(12):2831-2846.DOI:10.11834/jrs.20232333.
[17] TSAI F J, PENG Y T, LIN Y, et al.Stripformer:strip transformer for fast image deblurring[C]//Computer Vision- ECCV 2022:LNCS Volume 13679.Cham:Springer Nature Switzerland AG, 2022:146-162.DOI:10.1007/978-3-031-19800-7_9.
[18] 马六, 毛克彪, 郭中华.基于混合注意力生成对抗网络的遥感图像去雾方法[J].智慧农业(中英文), 2025, 7(2):172-182.DOI:10.12133/j.smartag.SA202410011.
[19] CAI B L, XU X M, JIA K, et al.DehazeNet:an end-to-end system for single image haze removal[J].IEEE Transactions on Image Processing, 2016, 25(11):5187-5198.DOI:10.1109/TIP.2016.2598681.
[20] 李愿, 付辉, 刘浩志.双重注意力下的多尺度残差遥感图像去雾网络[J].自然资源遥感, 2025, 37(4):31-39.DOI:10.6046/zrzyyg.2024154.
[21] WANG Z D, CUN X D, BAO J M, et al.Uformer:a general U-shaped transformer for image restoration[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Los Alamitos, CA:IEEE Computer Society, 2022:17662-17672.DOI:10.1109/CVPR52688.2022.01716.
[22] 杨燕, 晁丽鹏.基于多维协同注意力的双支特征联合去雾网络[J].浙江大学学报(工学版), 2025, 59(6):1119-1129.DOI:10.3785/j.issn.1008-973X.2025.06.003.
[23] KULKARNI A, MURALA S.Aerial image dehazing with attentive deformable transformers[C]//2023 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV).Los Alamitos, CA:IEEE Computer Society, 2023:6294-6303.DOI:10.1109/WACV56688.2023.00624.
[24] 周景, 田兆星, 王满意.融合外部注意力的扩散模型巡检图像去雾[J].电子测量技术, 2024, 47(15):144-152.DOI:10.19651/j.cnki.emt.2416080.
[25] YU H, ZHENG N S, ZHOU M, et al.Frequency and spatial dual guidance for image dehazing[C]//Computer Vision - ECCV 2022:LNCS Volume 13679.Cham:Springer Nature Switzerland AG, 2022:181-198.DOI:10.1007/978-3-031-19800-7_11.
[26] KHMAG A.Image dehazing and defogging based on second-generation wavelets and estimation of transmission map[J].Multimedia Tools and Applications, 2024, 83(19):57089-57105.DOI:10.1007/s11042-023-17819-z.
[27] JIANG B, CHEN G T, WANG J S, et al.Deep dehazing network for remote sensing image with non-uniform haze[J].Remote Sensing, 2021, 13(21):4443.DOI:10.3390/rs13214443.
[28] LI Y, ZHAO Y F.RSID:a remote sensing image dehazing network[C]//Pattern Recognition and Computer Vision:LNCS Volume 14430.Singapore:Springer Nature Singapore Pte Ltd., 2024:3-14.DOI:10.1007/978-981-99-8537-1_1.
[29] ZAMIR S W, ARORA A, KHAN S, et al.Restormer:efficient transformer for high-resolution image restoration[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Los Alamitos, CA:IEEE Computer Society, 2022:5718-5729.DOI:10.1109/CVPR52688.2022.00564.
[30] NUTHI R, KANKANALA S.SWAM-Net+:selective wavelet attentive M-network+ for single image dehazing[J].Circuits Systems and Signal Processing, 2025, 44(1):281-305.DOI:10.1007/s00034-024-02837-5.
[31] KONG L S, DONG J X, GE J J, et al.Efficient frequency domain-based transformers for high-quality image deblurring[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Los Alamitos, CA:IEEE Computer Society, 2022:5886-5895.DOI:10.1109/CVPR52729.2023.00570.
[32] DWIVEDI P, CHAKRABORTY S.FrTrGAN:single image dehazing using the frequency component of transmission maps in the generative adversarial network[J].Computer Vision and Image Understanding, 2025, 255:104336.DOI:10.1016/j.cviu.2025.104336.
[33] 刘春黔, 林浩然, 雷印杰.面向多浓度非均匀云雾的双域遥感图像去雾算法[J].现代计算机, 2024, 30(3):9-17.DOI:10.3969/j.issn.1007-1423.2024.03.002.
[34] HUANG B H, LI Z, YANG C, et al.Single satellite optical imagery dehazing using SAR image prior based on conditional generative adversarial networks[C]//2020 IEEE Winter Conference on Applications of Computer Vision (WACV).Los Alamitos, CA:IEEE Computer Society, 2020:1795-1802.DOI:10.1109/WACV45572.2020.9093471.
[35] LIN D Y, XU G L, WANG X K, et al.A remote sensing image dataset for cloud removal[EB/OL].(2019-01-03)[2025-06-10].https://arxiv.org/abs/1901.00600.DOI:10.48550/arXiv.1901.00600.
[36] FU M H, LIU H, YU Y K, et al.DW-GAN:a discrete wavelet transform GAN for nonhomogeneous dehazing[C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).Los Alamitos, CA:IEEE Computer Society, 2021:203-212.DOI:10.1109/CVPRW53098.2021.00029.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed