自适应归一化的多尺度水下图像增强网络

doi:10.16088/j.issn.1001-6600.2025071501

摘要/Abstract

摘要： 水下图像常因复杂环境导致颜色失真与细节丢失,现有图像增强方法通常对色彩通道统一处理,忽略其衰减差异,且Transformer受窗口分割机制限制,实际感受野较小,制约全局信息建模。为此,本文提出一种自适应归一化的多尺度水下图像增强网络。该方法融合卷积的局部特征提取优势与多尺度注意力的全局建模能力,采用两阶段架构:第1阶段通过多分支结构进行多尺度特征提取,捕获丰富的上下文信息;第2阶段基于编解码结构进行特征增强与重建,引入动态特征增强模块以自适应强化关键特征,并采用并行跳跃连接,有效提升特征完整性与结构一致性,最大化信息利用率。在UIEB、EUVP和LSUI公开数据集上的实验表明,本文方法PSNR/SSIM指标分别达到23.32/0.91、27.16/0.89和25.82/0.88,在定量与定性评估中均显著优于多种先进方法,有效改善颜色失真与细节模糊问题,提升整体视觉质量。

关键词: 水下图像增强, 深度学习, 多尺度, 自适应归一化, 色彩空间

Abstract: Underwater images often suffer from color distortion and detail loss due to complex environments. Existing methods process color channels uniformly, ignoring their distinct characteristics, while Transformers underperform convolutional networks due to limited information utilization. To address these issues, a multi-scale attention and adaptive normalization-based underwater image enhancement network is proposed. The network consists of two stages: multi-scale feature extraction and feature enhancement with reconstruction. In the first stage, rich features are captured through multi-scale processing. In the second stage, an encoder-decoder structure is employed, incorporating a feature enhancement module and parallel skip connections to ensure feature integrity and structural consistency while maximizing information utilization. Experimental results demonstrate that the proposed network significantly improves color correction and detail preservation compared with existing methods, achieving superior qualitative and quantitative performance.

Key words: underwater image enhancement, deep learning, multi-scale, adaptive normalization, color space

中图分类号: TP391.41

王燕, 徐婕, 牛梦圆. 自适应归一化的多尺度水下图像增强网络[J]. 广西师范大学学报（自然科学版）, 2026, 44(3): 89-106.

WANG Yan, XU Jie, NIU Mengyuan. Multi-scale Underwater Image Enhancement Network with Adaptive Normalization[J]. Journal of Guangxi Normal University(Natural Science Edition), 2026, 44(3): 89-106.

参考文献

[1] 李加林, 沈满洪, 马仁锋, 等. 海洋生态文明建设背景下的海洋资源经济与海洋战略[J]. 自然资源学报, 2022, 37(4): 829-849. DOI: 10.31497/zrzyxb.20220401.
[2] 李哲怡. 水下图像恢复算法综述[J]. 图像与信号处理, 2025, 14(2): 257-270. DOI: 10.12677/jisp.2025.142024.
[3] 严浙平, 曲思瑜, 邢文. 水下图像增强方法研究综述[J]. 智能系统学报, 2022, 17(5): 860-873. DOI: 10.11992/tis.202108022.
[4] 米泽田, 晋洁, 李圆圆, 等. 基于多尺度级联网络的水下图像增强方法[J]. 电子与信息学报, 2022, 44(10): 3353-3362. DOI: 10.11999/JEIT220375.
[5] 纪勋, 冷娜, 郭慧. 水下图像增强与复原技术进展与展望[J]. 计算机辅助设计与图形学学报, 2024, 36(6): 805-830. DOI: 10.3724/SP.J.1089.2024.20116.
[6] 周飞燕, 金林鹏, 董军. 卷积神经网络研究综述[J]. 计算机学报, 2017, 40(6): 1229-1251. DOI: 10.11897/SP.J.1016.2017.01229.
[7] LI C Y, ANWAR S, PORIKLI F. Underwater scene prior inspired deep underwater image and video enhancement[J]. Pattern Recognition, 2020, 98: 107038. DOI: 10.1016/j.patcog.2019.107038.
[8] WANG Y D, GUO J C, GAO H, et al. UEC²-Net: CNN-based underwater image enhancement using two color space[J]. Signal Processing: Image Communication, 2021, 96: 116250. DOI: 10.1016/j.image.2021.116250.
[9] LI J, SKINNER K A, EUSTICE R M, et al. WaterGAN:unsupervised generative network to enable real-time color correction of monocular underwater images[J]. IEEE Robotics and Automation Letters, 2018, 3(1): 387-394. DOI: 10.1109/LRA.2017.2730363.
[10] ISLAM M J, XIA Y Y, SATTAR J. Fast underwater image enhancement for improved visual perception[J]. IEEE Robotics and Automation Letters, 2020, 5(2): 3227-3234. DOI: 10.1109/LRA.2020.2974710.
[11] ZHOU T, LI Q, LU H L, et al. GAN review: models and medical image fusion applications[J]. Information Fusion, 2023, 91: 134-148. DOI: 10.1016/j.inffus.2022.10.017.
[12] 梁俊杰, 韦舰晶, 蒋正锋. 生成对抗网络GAN综述[J]. 计算机科学与探索, 2020, 14(1): 1-17.
[13] 丛晓峰, 桂杰, 章军. 基于视觉Transformer的多损失融合水下图像增强网络[J]. 智能科学与技术学报, 2022, 4(4):522-532. DOI: 10.11959/j.issn.2096-6652.202252.
[14] PENG L T, ZHU C L, BIAN L H. U-shape transformer for underwater image enhancement[J]. IEEE Transactions on Image Processing, 2023, 32: 3066-3079. DOI: 10.1109/TIP.2023.3276332.
[15] 刘文婷, 卢新明. 基于计算机视觉的Transformer研究进展[J]. 计算机工程与应用, 2022, 58(6): 1-16. DOI: 10.3778/j.issn.1002-8331.2106-0442.
[16] 郭佳霖, 智敏, 殷雁君, 等. 图像处理中CNN与视觉Transformer混合模型研究综述[J]. 计算机科学与探索, 2025, 19(1): 30-44. DOI: 10.3778/j.issn.1673-9418.2403009.
[17] YAO H Y, GUI R G, ZHAO Z D, et al. U-TransCNN: a U-shape transformer-CNN fusion model for underwater image enhancement[J]. Displays, 2025, 88: 103047. DOI: 10.1016/j.displa.2025.103047.
[18] CHEN X Y, WANG X T, ZHANG W L, et al. HAT: hybrid attention transformer for image restoration[EB/OL]. (2023-09-11)[2025-07-15]. https://arxiv.org/abs/2309.05239. DOI: 10.48550/arXiv.2309.05239.
[19] GU J J, DONG C. Interpreting super-resolution networks with local attribution maps[C]// 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2021: 9195-9204. DOI: 10.1109/CVPR46437.2021.00908.
[20] GU J, YE J C. AdaIN-based tunable CycleGAN for efficient unsupervised low-dose CT denoising[J]. IEEE Transactions on Computational Imaging, 2021, 7: 73-85. DOI: 10.1109/TCI.2021.3050266.
[21] CARRASCO M. How visual spatial attention alters perception[J]. Cognitive Processing, 2018, 19(1): 77-88. DOI: 10.1007/s10339-018-0883-4.
[22] HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Los Alamitos, CA: IEEE Computer Society, 2018: 7132-7141. DOI: 10.1109/CVPR.2018.00745.
[23] LI C Y, GUO C L, REN W Q, et al. An underwater image enhancement benchmark dataset and beyond[J]. IEEE Transactions on Image Processing, 2019, 29: 4376-4389. DOI: 10.1109/TIP.2019.2955241.
[24] BERMAN D, LEVY D, AVIDAN S, et al. Underwater single image color restoration using haze-lines and a new quantitative dataset[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 43(8): 2822-2837. DOI: 10.1109/TPAMI.2020.2977624.
[25] SIMOES W, DE SÁ M. PSNR and SSIM:evaluation of the imperceptibility quality of images transmitted over wireless networks[J]. Procedia Computer Science, 2024, 251: 463-470. DOI: 10.1016/j.procs.2024.11.134.
[26] KUFA J, BUDAC A. Quality comparison of 360° 8K images compressed by conventional and deep learning algorithms[C]// 2023 33rd International Conference Radioelektronika (RADIOELEKTRONIKA). Piscataway, NJ: IEEE, 2023: 1-4. DOI: 10.1109/RADIOELEKTRONIKA57919.2023.10109066.
[27] HIMANSHU, HOODA R, POPLY V. Image enhancement in Fourier domain using MSE and PSNR[J]. AIP Conference Proceedings. 2024, 3081(1): 060001. DOI: 10.1063/5.0196118.
[28] GHAZANFARI S, GARG S, KRISHNAMURTHY P, et al. R-LPIPS:an adversarially robust perceptual similarity metric[EB/OL]. (2023-07-31)[2025-07-15]. https://arxiv.org/abs/2307.15157. DOI: 10.48550/arXiv.2307.15157.
[29] LIU Y T, GU K, CAO J C, et al. UIQI: a comprehensive quality evaluation index for underwater images[J]. IEEE Transactions on Multimedia, 2023, 26: 2560-2573. DOI: 10.1109/TMM.2023.3301226.
[30] SINGH N, BHAT A. A robust model for improving the quality of underwater images using enhancement techniques[J]. Multimedia Tools and Applications, 2024, 83(1): 2267-2288. DOI: 10.1007/s11042-023-15617-1.
[31] MITTAL A, SOUNDARARAJAN R, BOVIK A C. Making a "completely blind" image quality analyzer[J]. IEEE Signal Processing Letters, 2013, 20(3): 209-212. DOI: 10.1109/LSP.2012.2227726.
[32] PENG Y T, COSMAN P C. Underwater image restoration based on image blurriness and light absorption[J]. IEEE Transactions on Image Processing, 2017, 26(4): 1579-1594. DOI: 10.1109/TIP.2017.2663846.
[33] XIAO Z Y, HAN Y N, RAHARDJA S, et al. USLN: a statistically guided lightweight network for underwater image enhancement via dual-statistic white balance and multi-color space stretch[EB/OL]. (2022-09-06)[2025-07-15]. https://arxiv.org/abs/2209.02221. DOI: 10.48550/arXiv.2209.02221.
[34] ZHANG W D, ZHUANG P X, SUN H H, et al. Underwater image enhancement via minimal color loss and locally adaptive contrast enhancement[J].IEEE Transactions on Image Processing, 2022, 31: 3997-4010. DOI: 10.1109/TIP.2022.3177129.
[35] FU Z Q, WANG W, HUANG Y, et al. Uncertainty inspired underwater image enhancement[C]// Computer Vision-ECCV 2022. Cham: Springer Nature Switzerland, 2022: 465-482. DOI: 10.1007/978-3-031-19797-0_27.
[36] ZHANG W D, JIN S L, ZHUANG P X, et al. Underwater image enhancement via piecewise color correction and dual prior optimized contrast enhancement[J]. IEEE Signal Processing Letters, 2023, 30: 229-233. DOI: 10.1109/LSP.2023.3255005.
[37] SHEN Z, XU H Y, LUO T, et al. UDAformer: underwater image enhancement based on dual attention transformer[J]. Computers & Graphics, 2023, 111: 77-88. DOI: 10.1016/j.cag.2023.01.009.
[38] HUANG Z X, LI J J, WANG X Y, et al. Underwater image enhancement via cross-wise transformer network focusing on pre-post differences[J]. Applied Soft Computing, 2024, 164: 112000. DOI: 10.1016/j.asoc.2024.112000.
[39] ZHANG S, LI D L, ZHAO R. PDCFNet: enhancing underwater images through pixel difference convolution[EB/OL]. (2024-09-28)[2025-07-15]. https://arxiv.org/abs/2409.19269. DOI: 10.48550/arXiv.2409.19269.
[40] ZHAO C, CAI W L, DONG C Y, et al. Toward sufficient spatial-frequency interaction for gradient-aware underwater image enhancement[C]// ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2024: 3220-3224. DOI: 10.1109/ICASSP48485.2024.10448182.
[41] XU C P, ZHOU W H, HUANG Z X, et al. Fusion-based graph neural networks for synergistic underwater image enhancement[J]. Information Fusion, 2025, 117: 102857. DOI: 10.1016/j.inffus.2024.102857.
[42] SALEH A, SHEAVES M, JERRY D, et al. Adaptive deep learning framework for robust unsupervised underwater image enhancement[J]. Expert Systems with Applications, 2025, 268: 126314. DOI: 10.1016/j.eswa.2024.126314.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed