计算机科学

计算机科学 ›› 2026, Vol. 53 ›› Issue (7): 62-70.doi: 10.11896/jsjkx.250400138

• 计算机图形学 & 多媒体 • 上一篇    下一篇

基于多级光照激励与联合损失约束的低光图像增强网络

焦韩冰1, 康俊华1, 肖腾2,3, 邓非4   

  1. 1 长安大学地质工程与测绘学院 西安 710000
    2 湖北工业大学计算机学院 武汉 430068
    3 汉诺威大学摄影测量与地理信息所 德国 汉诺威 30167
    4 武汉大学测绘学院 武汉 430000
  • 收稿日期:2025-04-28 修回日期:2025-08-12 出版日期:2026-07-15 发布日期:2026-07-10
  • 通讯作者: 康俊华(junhua.kang@chd.edu.cn)
  • 作者简介:(2024126050@chd.edu.cn)
  • 基金资助:
    陕西省自然科学基础研究计划(2024JC-YBQN-0325);国家资助博士后研究人员计划(GZC20232219);长安大学中央高校基本科研业务费专项资金(300102264102)

Low-light Image Enhancement Network Based on Multi-level Illumination Excitation and JointLoss Constraint

JIAO Hanbing1, KANG Junhua1, XIAO Teng2,3, DENG Fei4   

  1. 1 School of Geological Engineering and Geomatics,Chang'an University,Xi'an 710000,China
    2 School of Computer Science,Hubei University of Technology,Wuhan 430068,China
    3 Institute of Photogrammetry and GeoInformation,Leibniz Universität Hannover,Hannover 30167,Germany
    4 School of Geodesy and Geomatics,Wuhan University,Wuhan 430000,China
  • Received:2025-04-28 Revised:2025-08-12 Published:2026-07-15 Online:2026-07-10
  • About author:JIAO Hanbing,born in 2002,postgra-duate.His main research interests include photogrammetry and remotesen-sing,computer vision,and so on.
    KANG Junhua,born in 1990,lecturer.Her main research interests include computer vision and 3D reconstruction of drone images.
  • Supported by:
    Shaanxi Provincial Natural Science Basic Research Program(2024JC-YBQN-0325),State-Funded Postdoctoral Researcher Program(GZC20232219) and Fundamental Research Funds of the Central Universities of Chang'an University(300102264102).

PDF (PC)

17

摘要： 在自动驾驶、同步建图与定位(SLAM)等计算机视觉应用中,低照度环境下图像常因对比度下降、噪声干扰和细节丢失而严重制约视觉系统感知性能。现有低光图像增强方法在噪声抑制、色彩保真等方面仍存在不足,且跨场景泛化能力薄弱。为此,提出一种基于改进Retinexformer架构的深度学习低光图像增强方法,通过多阶段特征激励、全局光照调整与多维约束优化策略实现对低光图像的有效增强。首先,构建多级光照特征激励模块(MIFIB),通过特征归一化与改进型通道注意力机制实现特征表达增强,增强模型对光照特征的建模能力;其次,设计具备全局感知的光照调整模块(IAB),有效优化增强图像的光照分布;最后,提出融合结构相似性约束、语义特征约束及色彩强度一致性约束的多维联合损失优化策略,指导模型学习,实现更全面的图像质量优化。实验结果表明,在LOL标准低光数据集上,所提方法在PSNR和SSIM等指标上均优于主流模型;在SYNTHIA和Terrasentia数据集上的泛化测试和可视化效果进一步验证了该方法在抑制噪声、保持色彩和细节方面的优势。此外,在基于立体匹配的三维重建定量评估中,该方法增强图像立体匹配结果的端点误差(EPE)降低0.5,相较于无图像增强方法提升22.1%,充分体现了其在保持场景几何一致性和提升深度感知鲁棒性方面的优势。

关键词: 低光图像增强, 特征挤压激励, 光照调整, 深度学习, 立体匹配

Abstract: In computer vision applications such as autonomous driving and simultaneous localization and mapping(SLAM),low-light images frequently suffer from diminished contrast,noise interference,and detail loss,significantly impairing visual perception systems.Existing low-light enhancement methods exhibit limitations in noise suppression and color fidelity while demonstrating weak cross-scenario generalization capabilities.To address these challenges,this paper proposes a deep learning-based low-light enhancement approach using an improved Retinexformer architecture.The proposed method achieves effective enhancement through multi-stage feature excitation,global illumination adjustment,and multi-dimensional constrained optimization strategies.Firstly,it constructs MIFIB(Multi-level Illumination Feature Incentive Block) that enhances feature representation through normalization and an advanced channel attention mechanism,strengthening illumination modeling.Secondly,it designs a globally-aware IAB(Illumination Adjustment Block) to optimize illumination distribution in enhanced images.Finally,it introduces a multi-dimensional joint loss optimization strategy incorporating structural similarity constraints,semantic feature constraints,and color intensity consistency constraints to comprehensively guide model learning.Experimental results demonstrate that the proposed method achieves superior performance over state-of-the-art methods on the LOL benchmark in metrics including PSNR and SSIM.Generalization tests on SYNTHIA and Terrasentia datasets further validate the proposed method's advantages in noise suppression,color preservation,and detail retention.Moreover,quantitative evaluation in stereo matching-based 3D reconstruction shows that the proposed method reduces endpoint error(EPE) by 0.5 pixels－a 22.1% improvement compared to non-enhanced methods－confirming the advantages of the proposed method in maintaining scene geometric consistency and improving the robustness of depth perception.

Key words: Low-light image enhancement, Feature squeeze & excitation, Illumination adjustment, Deep learning, Stereo matching

中图分类号: 

  • TP391.41
[1]LYU Z W,NIU H J,SUN F Y et al.Review of Research on Low-Light Image Enhancement Algorithms[J].Infrared Technology,2025,47(2):165-178.
[2]LAND E H.The retinex theory of color vision[J].Scientific American,1977,237(6):108-129.
[3]LIU S G,ZHANG Y.Detail-preserving underexposed image enhancement via optimal weighted multi-exposure fusion[J].IEEE Transactions on Consumer Electronics,2019,65(3):303-311.
[4]VASWANI A,SHAZEER N,PARMAR N,et al.Attention isall you need[C]//Proceedings of the 31st International Confe-rence on Neural Information Processing Systems.2017:6000-6010.
[5]XU X,WANG R,FU C W,et al.Snr-aware low-light image enhancement[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2022:17714-17724.
[6]CAI Y,BIAN H,LIN J,et al.Retinexformer:One-stage retinex-based transformer for low-light image enhancement[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision.2023:12504-12513.
[7]ABDULLAH-AL-WADUD M,KABIR M H,DEWAN M A A,et al.A dynamic histogram equalization for image contrast enhancement[J].IEEE Transactions on Consumer Electronics,2007,53(2):593-600.
[8]CELIK T,TJAHJADI T.Contextual and variational contrastenhancement[J].IEEE Transactions on Image Processing,2011,20(12):3431-3441.
[9]LI S,KANG X.Fast multi-exposure image fusion with median filter and recursive filter[J].IEEE Transactions on Consumer Electronics,2012,58(2):626-632.
[10]HUANG S C,CHENG F C,CHIU Y S.Efficient contrast enhancement using adaptive gamma correction with weighting distribution[J].IEEE Transactions on Image Processing,2012,22(3):1032-1041.
[11]LIU Y,JIA X F.Low-light image enhancement algorithm in-tegrating residual structure and attention mechanism[J].Journal of Chongqing Technology and Business University(Natural Science Edition),2024(4):86-96.
[12]GUO X J,LI Y,LING H.LIME:Low-light image enhancement via illumination map estimation[J].IEEE Transactions on Image Processing,2017,26:982-993.
[13]WANG D W,ZHAO W L,FANG J,et al.Low-light image en-hancement algorithm for UAV aerial photography base on low-light instances[J/OL].Journal of Beijing University of Aeronautics and Astronautics.DOI:10.13700/j.bh.1001-5965.2024.0531.
[14]WEI C,WANG W,YANG W,et al.Deep retinex decomposition for low-light enhancement[C]//Proceedings of the British Machine Vision Conference(BMVC).2018:1-12.
[15]ZHANG Y H,GUO X J,MA J Y,et al.Beyond brightening low-light images[J].International Journal of Computer Vision,2021,129(4):1013-1037.
[16]ZHANG Y,ZHANG J,GUO X.Kindling the darkness:A practical low-light image enhancer[C]//Proceedings of the 27th ACM International Conference on Multimedia.New York:ACM,2019:1632-1640.
[17]ZHAO L,LU S P,CHEN T,et al.Deep symmetric network for underexposed image enhancement with recurrent attentional learning[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision.Piscataway,NJ:IEEE,2021:12075-12084.
[18]WANG R,ZHANG Q,FU C W,et al.Underexposed photo enhancement using deep illumination estimation[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2019:6849-6857.
[19]ZHANG Z,JIANGY,JIANG J,et al.STAR:A structure-aware lightweight transformer for real-time image enhancement[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision.Piscataway,NJ:IEEE,2021:4106-4115.
[20]LIU B,TIAN G L,XIAO B,et al.Low Light Image Enhancement With Adaptive Light Initialization[J].Journal of Electro-nics & Information Technology,2024,46(2):643-651.
[21]HU J,SHEN L,SUN G.Squeeze-and-excitation networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Piscataway,NJ:IEEE,2018:7132-7141.
[22]BRATEANU A,BALMEZ R,AVRAM A,et al.LYT-Net:Lightweight YUV transformer-based network for low-light image enhancement[J].IEEE Signal Processing Letters,2025,32:2065-2069.
[23]WEI C,WANG W,YANG W,et al.Deep retinex decomposition for low-lightenhancement[J].arXiv:1808.04560,2018.
[24]YANG W,WANG W,HUANG H,et al.Sparse gradient regularized deep retinex networkfor robust low-light image enhancement[J].IEEE Transactions on Image Processing,2021,30:2072-2086.
[25]ROS G,SELLART L,MATERZYNSKA J,et al.The synthia dataset:A large collection of synthetic images for semantic segmentation of urban scenes[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2016:3234-3243.
[26]PASZKE A.Pytorch:An imperative style,high-performancedeep learning library[J].arXiv:1912.01703,2019.
[27]KINGMA D P,BA J.Adam:A method for stochastic optimization[J].arXiv:1412.6980,2014.
[28]LOSHCHILOV I,HUTTER F.Sgdr:Stochastic gradient de-scent with warm restarts[J].arXiv:1608.03983,2016.
[29]WANG Z,BOVIK A C,SHEIKH H R,et al.Image quality assessment:from error visibility to structural similarity[J].IEEE Transactions on Image Processing,2004,13(4):600-612.
[30]SUN F Y,LYU Z,LYU Z W.Review of low light image enhancement based on deep learning [J].Application Research of Computers,2025,42(1):19-27.
[31]YANG W,WANG S,FANG Y,et al.Band representation-based semi-supervised low-light image enhancement:Bridging the gap between signal fidelity and perceptual quality[J].IEEE Transactions on Image Processing,2021,30:3461-3473.
[32]ZAMIR S W,ARORA A,KHAN S,et al.Restormer:Efficient transformer for high-resolution image restoration[C]//Procee-dings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2022:5728-5739.
[33]CUI Z T,LI K C,GU L,et al.You only need 90K parameters to adapt light:A light weight transformer for image enhancement and exposure correction[C]//Proceedings of the 33rd British Machine Vision Conference.2022:1-12.
[34]WANG T,ZHANG K,SHEN T,et al.Ultra-high-definitionlow-light image enhancement:A benchmark and transformer-based method[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2023:2654-2662.
[35]LI C,GUO C L,ZHOU M,et al.Embedding fourier for ultra-high-definition low-light image enhancement[J].arXiv:2302.11831,2023.
[36]TAN J,PEI S,QIN W,et al.Wavelet-based Mamba with Fourier Adjustment for Low-light Image Enhancement[C]//Proceedings of the Asian Conference on Computer Vision.2024:3449-3464
[37]ZAMIR S W,ARORA A,KHAN S,et al.Learning enriched features for real image restoration and enhancement[C]//Compu-ter Vision-ECCV 2020:16th European Conference.Springer,2020:492-511.
[38]XU G,WANG X,ZHANG Z,et al.IGEV++:Iterative multi-range geometry encoding volumes for stereo matching[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2025,47(8):7108-7122.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市两江新区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发