Underwater image enhancement algorithm based on multi-scale perception and multi-dimensional space fusion

doi:10.11772/j.issn.1001-9081.2025010139

Journal of Computer Applications ›› 2026, Vol. 46 ›› Issue (1): 224-232.DOI: 10.11772/j.issn.1001-9081.2025010139

• Multimedia computing and computer simulation • Previous Articles Next Articles

Underwater image enhancement algorithm based on multi-scale perception and multi-dimensional space fusion

Wei GUO, Manting WANG(), Haicheng QU

School of Software，Liaoning Technical University，Huludao Liaoning 125105，China

Received:2025-02-13 Revised:2025-04-07 Accepted:2025-04-08 Online:2026-01-10 Published:2026-01-10
Contact: Manting WANG
About author:GUO Wei， born in 1970， M. S.， associate professor. Her research interests include computer vision， intelligent image processing.
QU Haicheng， born in 1981， Ph. D.， associate professor. His research interests include intelligent processing of remote sensing big data， target recognition and tracking.
Supported by:
National Natural Science Foundation of China(42271409);Basic Scientific Research Project of Higher Education Institutions of Liaoning Province(LJKMZ20220699)

基于多尺度感知的多维空间融合水下图像增强算法

郭伟, 王曼婷(), 曲海成

辽宁工程技术大学软件学院，辽宁葫芦岛 125105

通讯作者: 王曼婷
作者简介:郭伟（1970—），女，辽宁阜新人，副教授，硕士， CCF会员，主要研究方向：计算机视觉、智能图像处理
曲海成（1981—），男，山东烟台人，副教授，博士， CCF会员，主要研究方向：遥感大数据智能处理、目标识别与跟踪。
基金资助:
国家自然科学基金资助项目(42271409);辽宁省高等学校基本科研项目(LJKMZ20220699)

Abstract

Abstract:

To address problems caused by deep-sea imaging， such as color distortion， low contrast， and blurred structures in underwater images， an underwater image enhancement algorithm based on multi-scale perception and multi-dimensional space fusion was proposed. By combining spatial， channel， and three-dimensional features， image information was transmitted in parallel by the algorithm to a multi-dimensional feature extraction network and an encoder. Firstly， a multiscale feature refinement module was introduced into the feature extraction network to further process the extracted feature information， allowing the network to learn information at different scales more accurately. Secondly， a multidimensional color enhancement module was incorporated into the encoder to enhance image details and colors. Finally， an adaptive enhancement network was designed to further process the feature information and fuse multi-level features， then the decoder was used to generate the final enhanced image. Experimental results on public datasets demonstrate the outstanding performance of the proposed algorithm. Specifically， it achieves a Peak Signal-to-Noise Ratio （PSNR） of up to 24.865 1 dB and a Structural Similarity （SSIM） of 0.895 4， representing improvements of 1.580 6 dB and 0.039 8 over Hybrid Fusion Method （HFM）， respectively， and it has the Underwater Color Image Quality Evaluation （UCIQE） and Underwater Image Quality Measure （UIQM） up to 0.593 1 and 3.102 8， respectively， surpassing HFM by 0.038 4 and 0.151 4， respectively. It can be seen that the proposed algorithm improves underwater visual quality effectively.

Key words: image processing, feature extraction, multiscale feature, deep convolution, enhanced color

摘要：

针对深海拍摄会导致水下图像色彩偏移、对比度过低和结构不清晰等问题，提出一种基于多尺度感知的多维空间融合水下图像增强算法，结合空间、通道和三维特征将图像信息并行传入多维特征提取网络和编码器中。首先，在多维特征提取网络中引入多尺度特征精炼模块进一步处理提取到的特征信息，使网络更准确地学习不同尺度的信息；然后，在编码器中引入多维色彩增强模块，增强图像细节和色彩；最后，设计自适应增强网络来进一步处理特征信息并融合多级信息，再通过解码器得到最终的增强图像。在公开数据集上的实验结果表明，所提算法表现优异，它的峰值信噪比（PSNR）和结构相似性（SSIM）最高分别达到24.865 1 dB和0.895 4，比混合融合方法（HFM）分别提升了1.580 6 dB和0.039 8；水下色彩质量评价（UCIQE）和水下图像质量测量（UIQM）最高分别达到0.593 1和3.102 8，比HFM分别提升了0.038 4和0.151 4。可见，所提算法能有效提升水下视觉效果。

关键词: 图像处理, 特征提取, 多尺度特征, 深层卷积, 强化色彩

CLC Number:

TP391.41

Wei GUO, Manting WANG, Haicheng QU. Underwater image enhancement algorithm based on multi-scale perception and multi-dimensional space fusion[J]. Journal of Computer Applications, 2026, 46(1): 224-232.

郭伟, 王曼婷, 曲海成. 基于多尺度感知的多维空间融合水下图像增强算法[J]. 《计算机应用》唯一官方网站, 2026, 46(1): 224-232.

Figures/Tables 14

References 29

[1]	李大海，李冰涛，王振东.基于改进YOLOv8的水下目标检测算法［J］.计算机应用， 2024， 44（11）： 3610-3616.
	LI D H， LI B T， WANG Z D. Underwater target detection algorithm based on improved YOLOv8 ［J］. Journal of Computer Applications， 2024， 44（11）： 3610-3616.
[2]	HAN M， LYU Z， QIU T， et al. A review on intelligence dehazing and color restoration for underwater images ［J］. IEEE Transactions on Systems， Man， and Cybernetics： Systems， 2020， 50（5）： 1820-1832.
[3]	ANWAR S， LI C. Diving deeper into underwater image enhancement： a survey ［J］. Signal Processing： Image Communication， 2020， 89： No.115978.
[4]	LI C， ANWAR S， PORIKLI F. Underwater scene prior inspired deep underwater image and video enhancement ［J］. Pattern Recognition， 2020， 98： No.107038.
[5]	REZA A M. Realization of the Contrast Limited Adaptive Histogram Equalization （CLAHE） for real-time image enhancement ［J］. Journal of VLSI Signal Processing Systems for Signal， Image and Video Technology， 2004， 38（1）： 35-44.
[6]	GONZALEZ R C， WOODS R E. Digital image processing ［M］. London： Pearson， 2018.
[7]	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.
[8]	CARLEVARIS-BIANCO N， MOHAN A， EUSTICE R M. Initial results in underwater single image dehazing ［C］// Proceedings of the OCEANS 2010 MTS/IEEE SEATTLE. Piscataway： IEEE， 2010： 1-8.
[9]	HUANG D， WANG Y， SONG W， et al. Shallow-water image enhancement using relative global histogram stretching based on adaptive parameter acquisition ［C］// Proceedings of the 2018 International Conference on Multimedia Modeling， LNCS 10704. Cham： Springer， 2018： 453-465.
[10]	李慧，贾炳志，王晨曦，等.基于Swin Transformer的生成对抗网络水下图像增强模型［J］.计算机应用， 2025， 45（5）： 1439-1446.
	LI H， JIA B Z， WANG C X， et al. Generative adversarial network underwater image enhancement model based on Swin Transformer ［J］. Journal of Computer Applications， 2025， 45（5）： 1439-1446.
[11]	ZHU J Y， PARK T， ISOLA P， et al. Unpaired image-to-image translation using cycle-consistent adversarial networks ［C］// Proceedings of the 2017 IEEE International Conference on Computer Vision. Piscataway： IEEE， 2017： 2242-2251.
[12]	WANG D， MA L， LIU R， et al. Semantic-aware texture-structure feature collaboration for underwater image enhancement ［C］// Proceedings of the 2022 International Conference on Robotics and Automation. Piscataway： IEEE， 2022： 4592-4598.
[13]	LI C， GUO C， REN W， et al. An underwater image enhancement benchmark dataset and beyond ［J］. IEEE Transactions on Image Processing， 2020， 29： 4376-4389.
[14]	PENG L， ZHU C， BIAN L. U-shape Transformer for underwater image enhancement ［J］. IEEE Transactions on Image Processing， 2023， 32： 3066-3079.
[15]	ZHANG Y， CHANDLER D M， LESZCZUK M. Retinex-based underwater image enhancement via adaptive color correction and hierarchical U-shape Transformer ［J］. Optics Express， 2024， 32（14）： 24018-24040.
[16]	WANG H， ZHANG W， BAI L， et al. Metalantis： a comprehensive underwater image enhancement framework ［J］. IEEE Transactions on Geoscience and Remote Sensing， 2024， 62： No.5618319.
[17]	AN S， XU L， DENG Z， et al. HFM： a hybrid fusion method for underwater image enhancement ［J］. Engineering Applications of Artificial Intelligence， 2024， 127（Pt A）： No.107219.
[18]	DING X， ZHANG X， MA N， et al. RepVGG： Making VGG-style ConvNets great again ［C］// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2021： 13728-13737.
[19]	DENG J， DONG W， SOCHER R， et al. ImageNet： a large-scale hierarchical image database ［C］// Proceedings of the 2009 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2009： 248-255.
[20]	WEBB B S， DHRUV N T， SOLOMON S G， et al. Early and late mechanisms of surround suppression in striate cortex of macaque ［J］. Journal of Neuroscience， 2005， 25（50）： 11666-11675.
[21]	LAI W S， HUANG J B， AHUJA N， et al. Fast and accurate image super-resolution with deep Laplacian pyramid networks ［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2019， 41（11）： 2599-2613.
[22]	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.
[23]	JOHNSON J， ALAHI A， LI F F. Perceptual losses for real-time style transfer and super-resolution ［C］// Proceedings of the 2016 European Conference on Computer Vision， LNCS 9906. Cham： Springer， 2016： 694-711.
[24]	ISLAM M J， XIA Y， SATTAR J. Fast underwater image enhancement for improved visual perception ［J］. IEEE Robotics and Automation Letters， 2020， 5（2）： 3227-3234.
[25]	ZHANG L， ER M J， WANG Z， et al. Underwater image enhancement based on AttR2U-Net and multi-residual networks ［C］// Proceedings of the 5th International Conference on Intelligent Autonomous Systems. Piscataway： IEEE， 2022： 90-95.
[26]	KORHONEN J， YOU J. Peak signal-to-noise ratio revisited： is simple beautiful？［C］// Proceedings of the 4th International Workshop on Quality of Multimedia Experience. Piscataway： IEEE， 2012： 37-38.
[27]	WANG Z， SIMONCELLI E P， BOVIK A C. Multi-scale structural similarity for image quality assessment ［C］// Proceedings of the 37th Asilomar Conference on Signals， Systems and Computers — Volume 2. Piscataway： IEEE， 2003： 1398-1402.
[28]	PANETTA K， GAO C， AGAIAN S. Human-visual-system-inspired underwater image quality measures ［J］. IEEE Journal of Oceanic Engineering， 2015， 41（3）： 541-551.
[29]	YANG M， SOWMYA A. An underwater color image quality evaluation metric ［J］. IEEE Transactions on Image Processing， 2015， 24（12）： 6062-6071.

数据集	训练集样本数	测试集样本数	合计
EUVP1	4 995	555	5 550
EUVP2	3 330	370	3 700
EUVP3	1 967	218	2 185
UIEB	801	89	890
LSUI	3 851	428	4 279

数据集	训练集样本数	测试集样本数	合计
EUVP1	4 995	555	5 550
EUVP2	3 330	370	3 700
EUVP3	1 967	218	2 185
UIEB	801	89	890
LSUI	3 851	428	4 279

算法	EUVP1		EUVP2		EUVP3
算法	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM
CLAHE^［5］	18.536 3	0.678 4	19.896 1	0.587 8	19.903 5	0.577 4
HE^［6］	14.632 5	0.654 8	15.651 5	0.612 4	14.854 9	0.569 6
IBLA^［7］	16.358 1	0.651 7	16.495 4	0.754 1	14.865 3	0.587 4
CycleGAN^［11］	19.102 8	0.768 9	20.633 2	0.720 2	22.402 3	0.789 4
AttR2U-Net^［25］	20.158 7	0.725 0	21.681 3	0.796 4	21.847 9	0.712 6
STSC^［12］	21.890 1	0.862 2	22.112 3	0.798 1	23.654 4	0.690 3
WaterNet^［13］	22.546 3	0.881 4	23.032 2	0.715 8	22.153 3	0.788 5
U-Shape^［14］	22.976 2	0.811 2	23.566 3	0.853 1	23.652 3	0.709 1
Metalantis^［16］	22.831 4	0.843 8	23.198 2	0.849 3	22.351 7	0.767 5
HFM^［17］	23.441 5	0.855 6	22.889 5	0.799 7	23.284 5	0.801 2
本文算法	23.033 9	0.895 4	24.655 4	0.866 4	24.865 1	0.813 9

算法	EUVP1		EUVP2		EUVP3
算法	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM
CLAHE^［5］	18.536 3	0.678 4	19.896 1	0.587 8	19.903 5	0.577 4
HE^［6］	14.632 5	0.654 8	15.651 5	0.612 4	14.854 9	0.569 6
IBLA^［7］	16.358 1	0.651 7	16.495 4	0.754 1	14.865 3	0.587 4
CycleGAN^［11］	19.102 8	0.768 9	20.633 2	0.720 2	22.402 3	0.789 4
AttR2U-Net^［25］	20.158 7	0.725 0	21.681 3	0.796 4	21.847 9	0.712 6
STSC^［12］	21.890 1	0.862 2	22.112 3	0.798 1	23.654 4	0.690 3
WaterNet^［13］	22.546 3	0.881 4	23.032 2	0.715 8	22.153 3	0.788 5
U-Shape^［14］	22.976 2	0.811 2	23.566 3	0.853 1	23.652 3	0.709 1
Metalantis^［16］	22.831 4	0.843 8	23.198 2	0.849 3	22.351 7	0.767 5
HFM^［17］	23.441 5	0.855 6	22.889 5	0.799 7	23.284 5	0.801 2
本文算法	23.033 9	0.895 4	24.655 4	0.866 4	24.865 1	0.813 9

算法	EUVP1		EUVP2		EUVP3
算法	UIQM	UCIQE	UIQM	UCIQE	UIQM	UCIQE
CLAHE^［5］	1.948 3	0.546 7	2.064 2	0.554 1	2.246 5	0.574 1
HE^［6］	2.036 4	0.511 2	1.956 4	0.541 2	2.345 8	0.535 4
IBLA^［7］	2.304 1	0.562 3	2.342 1	0.551 7	2.518 0	0.561 2
CycleGAN^［11］	2.742 9	0.556 4	2.982 9	0.523 1	2.839 4	0.603 1
AttR2U-Net^［25］	2.051 2	0.581 9	1.869 4	0.596 4	1.911 6	0.582 7
STSC^［12］	3.047 9	0.553 2	2.929 5	0.552 9	2.913 7	0.582 2
WaterNet^［13］	3.030 6	0.542 8	3.009 4	0.545 8	3.096 3	0.571 2
U-Shape^［14］	2.963 3	0.556 5	2.963 8	0.551 3	2.935 8	0.584 1
Metalantis^［16］	3.019 5	0.514 4	2.953 3	0.544 8	2.841 2	0.563 1
HFM^［17］	2.951 4	0.541 7	2.871 4	0.551 3	3.001 9	0.554 7
本文算法	3.102 8	0.580 1	3.015 4	0.563 9	3.056 7	0.593 1

Underwater image enhancement algorithm based on multi-scale perception and multi-dimensional space fusion

基于多尺度感知的多维空间融合水下图像增强算法

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 14

References 29

Related Articles 15

Recommended Articles

Metrics

算法	模型参数量/10⁶	算法运行时间/h	浮点运算量/GFLOPs
CLAHE^［5］	—	0.770 7	0.010
HE^［6］	—	0.735 0	0.005
IBLA^［7］	—	3.675 0	0.100
AttR2U-Net^［25］	—	10.634 0	25.000
CycleGAN^［11］	32.286 9	99.219 4	120.000
STSC^［12］	28.687 3	9.441 1	45.000
WaterNet^［13］	10.091 0	36.481 4	18.000
U-Shape^［14］	88.134 8	15.727 2	85.000
本文算法	86.372 1	6.583 5	40.000

模块	PSNR/dB	SSIM	UIQM	UCIQE	模型参数量/10⁶	算法运行时间/h
U	22.156 3	0.784 0	2.921 4	0.536 4	72.105 6	4.563 9
U+F	23.563 1	0.815 8	2.964 2	0.556 1	82.632 9	5.562 1
U+M	22.635 1	0.794 6	2.998 4	0.543 2	73.252 1	5.232 0
U+A	22.325 6	0.781 2	2.941 5	0.539 4	73.859 4	5.263 5
U+F+M	23.659 4	0.835 6	2.998 4	0.543 4	74.289 6	6.389 2
U+F+A	24.655 4	0.819 8	2.923 1	0.551 0	74.658 9	6.056 4
U+F+M+A	24.865 1	0.866 4	3.015 4	0.563 9	86.526 4	6.583 5

[1]	Xiuyan ZHANG, Wentao LIU, Xin WANG. Feature extraction method of flight pitch operation based on quick access recorder data [J]. Journal of Computer Applications, 2026, 46(1): 322-330.
[2]	Chuang WANG, Lu YU, Jianwei CHEN, Cheng PAN, Wenbo DU. Review of open set domain adaptation [J]. Journal of Computer Applications, 2025, 45(9): 2727-2736.
[3]	Jin XIE, Surong CHU, Yan QIANG, Juanjuan ZHAO, Hua ZHANG, Yong GAO. Dual-branch distribution consistency contrastive learning model for hard negative sample identification in chest X-rays [J]. Journal of Computer Applications, 2025, 45(7): 2369-2377.
[4]	Qingli CHEN, Yuanbo GUO, Chen FANG. Clustering federated learning algorithm for heterogeneous data [J]. Journal of Computer Applications, 2025, 45(4): 1086-1094.
[5]	Yu WANG, Xianjin FANG, Gaoming YANG, Yifeng DING, Xinlu YANG. Active defense against face forgery based on attention mask and feature extraction [J]. Journal of Computer Applications, 2025, 45(3): 904-910.
[6]	Tianqi ZHANG, Shuang TAN, Xiwen SHEN, Juan TANG. Image watermarking method combining attention mechanism and multi-scale feature [J]. Journal of Computer Applications, 2025, 45(2): 616-623.
[7]	Qiurun HE, Jie HU, Bo PENG, Tianyuan LI. Fabric defect detection algorithm based on context information and multi-scale feature fusion [J]. Journal of Computer Applications, 2025, 45(2): 640-646.
[8]	Jing HUANG, Xin PENG, Wenhao LI, Kai HU, Teng WANG, Yamin HUANG, Yuanqiao WEN. High-quality sonar image generation method based on multi-scale feature fusion [J]. Journal of Computer Applications, 2025, 45(12): 3987-3994.
[9]	Xiaoman WANG, Yanping CHEN, Caiwei YANG, Ruizhang HUANG, Yongbin QIN. Nested named entity recognition method for multi-directional gradient feature extraction [J]. Journal of Computer Applications, 2025, 45(11): 3547-3554.
[10]	Bingjie QIU, Chaoqun ZHANG, Weidong TANG, Bicheng LIANG, Danyang CUI, Haisheng LUO, Qiming CHEN. Zero-shot relation extraction model based on dual contrastive learning [J]. Journal of Computer Applications, 2025, 45(11): 3555-3563.
[11]	Lianqing WEN, Ye TAO, Yunlong TIAN, Li NIU, Hongxia SUN. Flow-based lightweight high-quality text-to-speech conversion method [J]. Journal of Computer Applications, 2025, 45(10): 3277-3283.
[12]	Pengcheng SONG, Lijun GUO, Rong ZHANG. Weakly supervised video anomaly detection with local-global temporal dependency [J]. Journal of Computer Applications, 2025, 45(1): 240-246.
[13]	Jietao LIANG, Bing LUO, Lanhui FU, Qingling CHANG, Nannan LI, Ningbo YI, Qi FENG, Xin HE, Fuqin DENG. Point cloud registration method based on coordinate geometric sampling [J]. Journal of Computer Applications, 2025, 45(1): 214-222.
[14]	Xin YANG, Xueni CHEN, Chunjiang WU, Shijie ZHOU. Short-term traffic flow prediction of urban highway based on variant residual model and Transformer [J]. Journal of Computer Applications, 2024, 44(9): 2947-2951.
[15]	Shuai FU, Xiaoying GUO, Ruyi BAI, Tao YAN, Bin CHEN. Age estimation method combining improved CloFormer model and ordinal regression [J]. Journal of Computer Applications, 2024, 44(8): 2372-2380.