基于语义传输的智能水下监测机制

doi:10.12305/j.issn.1001-506X.2026.09.36

系统工程与电子技术 ›› 2026, Vol. 48 ›› Issue (9): 3241-3250.doi: 10.12305/j.issn.1001-506X.2026.09.36

• 通信与网络 • 上一篇

基于语义传输的智能水下监测机制

金志刚, 常利祥(), 程思萌, 李根

天津大学电气自动化与信息工程学院，天津 300072

收稿日期:2025-08-22 修回日期:2025-11-07 出版日期:2026-01-24 发布日期:2026-01-24
通讯作者: 李根 E-mail:984602970@qq.com
作者简介:金志刚（1972—），男，教授，博士，主要研究方向为水下网络、传感器网络、网络安全、社交网络、大数据
常利祥（2001—），男，硕士研究生，主要研究方向为深度学习、水声通信
程思萌（1997—），女，博士研究生，主要研究方向为水下图像处理、水下图像传输
基金资助:
国家自然科学基金（52471364, 52171337）资助课题

Intelligent underwater monitoring mechanism based on semantic transmission

Zhigang Jin, Lixiang Chang(), Simeng Cheng, Gen Li

School of Electrical and Information Engineering，Tianjin University，Tianjin 300072，China

Received:2025-08-22 Revised:2025-11-07 Online:2026-01-24 Published:2026-01-24
Contact: Gen Li E-mail:984602970@qq.com

摘要/Abstract

摘要：

水声信道严重的噪声干扰和窄带宽特性限制了水下监测数据的传输，制约了水下监测的效率和智能化水平。基于此，提出一种基于语义传输的智能水下监测机制，设计了基于互信息变分近似的损失函数，整合图像重建失真与监测任务相关语义失真；构建多尺度特征融合编码器，融合局部与全局语义特征，并通过语义压缩模块保留关键语义特征以提升监测任务性能。实验结果表明，所提方法在保持一定图像重建质量的同时，在珊瑚礁分类与水下生物目标检测任务中性能更优，满足了水下监测的可视化需求，提高了水下监测的性能和智能化水平。

关键词: 水下监测机制, 语义传输, 监测性能, 图像传输

Abstract:

The severe noise interference and narrow bandwidth characteristics of the underwater acoustic channel limit the transmission of underwater monitoring data, thereby restricting the efficiency and level of intelligence of underwater monitoring. To deal with this, an intelligent underwater monitoring mechanism based on semantic transmission is proposed. A loss function based on the variational approximation of mutual information is designed to integrate the distortion of image reconstruction with the semantic distortion related to the monitoring task. A multi-scale feature fusion encoder is constructed to integrate local and global semantic features, and a semantic compression module retains key semantic features to enhance the performance of the monitoring task. Experimental results show that the proposed method achieves better performance in coral reef classification and underwater object detection tasks while maintaining a certain level of image reconstruction quality. It not only meets the visualization requirements of underwater monitoring but also improves the performance and level of intelligence of underwater monitoring.

Key words: underwater monitoring mechanism, semantic transmission, monitoring performance, image transmission

中图分类号:

TN 929.3

金志刚, 常利祥, 程思萌, 李根. 基于语义传输的智能水下监测机制[J]. 系统工程与电子技术, 2026, 48(9): 3241-3250.

Zhigang Jin, Lixiang Chang, Simeng Cheng, Gen Li. Intelligent underwater monitoring mechanism based on semantic transmission[J]. Systems Engineering and Electronics, 2026, 48(9): 3241-3250.

图/表 13

图1

图2

图3

表1

图4

图5

图6

图7

表2

图8

表3

表4

图9

参考文献 30

1	赵矣昊, 陈友淦, 李姜辉, 等. 人工智能技术在水声网络路由协议中的应用探索[J]. 电子与信息学报, 2025, 47 (8): 2429. doi: 10.11999/JEIT250110
2	金志刚, 段晨旭, 羊秋玲, 等. 基于水下云边协同架构的珊瑚礁监测新机制[J]. 系统工程与电子技术, 2022, 44 (12): 3829. doi: 10.12305/j.issn.1001-506X.2022.12.28
3	Qin Z J, Tao X M, Li J H, et al. Semantic communications: principles and challenges[EB/OL]. [2025-06-22]. https: //arxiv.org/abs/2201.01389.
4	张平, 牛凯, 姚圣时, 等. 面向未来的语义通信: 基本原理与实现方法[J]. 通信学报, 2023, 44 (5): 1. doi: 10.11959/j.issn.1000?436x.2023079
5	Zhang J S, Sun W K, Zhao Y K, et al. Semantic communication in underwater communication: advantage, problem and solution — a survey[C]// Proc. of the International Conference on Intelligent Computing and Signal Processing, 2023: 2120.
6	Picano B, Fantacci R. A semantic-oriented federated learning for hybrid ground-aqua computing systems[J]. IEEE Internet of Things Journal, 2024, 11 (6): 10095. doi: 10.1109/JIOT.2023.3325289
7	Yu L S, Zhang H J, Liu J L, et al. VimSC: robust underwater acoustic image semantic communication based on vision Mamba model[C]// Proc. of the International Conference on Communications and Broadband Networking, 2024: 46.
8	Xu Z P, Xu D Y, Lin L X, et al. Integrated object detection and communication for synthetic aperture radar images[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2025, 18, 294. doi: 10.1109/JSTARS.2024.3495023
9	Deng Z X, Zhang Y, Yang H B, et al. Underwater point cloud transmission framework: hybrid encoder implementation based on CNN and transformer[J]. Measurement Science and Technology, 2024, 36 (1): 015111. doi: 10.1088/1361-6501/ad8776
10	Jiang X, Zhang Y X, Pan M, et al. A marine organism detection framework based on dataset augmentation and CNN-ViT fusion[J]. Journal of Marine Science and Engineering, 2023, 11 (4): 705. doi: 10.3390/jmse11040705
11	Zhang Z D, Han Q L, Liu W P, et al. A lightweight network based on SCMYOLO for accurate and efficient underwater fish detection[J]. ICES Journal of Marine Science, 2025, 82 (4): 1021. doi: 10.1093/icesjms/fsaf038
12	Tong W J, Liu F F, Sun Z F, et al. Image semantic communications: an extended rate-distortion theory based scheme[C]// Proc. of the IEEE Globecom Workshops, 2022: 1723.
13	Huang J H, Li D X, Huang C, et al. Joint task and data-oriented semantic communications: a deep separate source-channel coding scheme[J]. IEEE Internet of Things Journal, 2024, 11 (2): 2255. doi: 10.1109/JIOT.2023.3293154
14	Berman D, Levy D, Avidan S, et al. Underwater single image color restoration using haze-lines and a new quantitative dataset[J]. IEEE Trans. on Pattern Analysis and Machine Intelligence, 2021, 43 (8): 2822. doi: 10.1109/tpami.2020.2977624
15	Zhang W D, Chen G C, Zhuang P X, et al. CATNet: cascaded attention transformer network for marine species image classification[J]. Expert Systems with Applications, 2024, 256, 124932. doi: 10.1016/j.eswa.2024.124932
16	Qu H S, Wang G G, Li Y, et al. ConvFishNet: an efficient backbone for fish classification from composited underwater images[J]. Information Sciences, 2024, 679, 121078. doi: 10.1016/j.ins.2024.121078
17	Leng Z Q, Tan M X, Liu C X, et al. PolyLoss: a polynomial expansion perspective of classification loss functions[C]// Proc. of the International Conference on Learning Representations, 2022.
18	Qarabaqi P, Stojanovic M. Statistical characterization and computationally efficient modeling of a class of underwater acoustic communication channels[J]. IEEE Journal of Oceanic Engineering, 2013, 38 (4): 701. doi: 10.1109/JOE.2013.2278787
19	Wang S, Chen N L, Song Y D, et al. ML-Net: a multi-local perception network for healthy and bleached coral image classification[J]. Journal of Marine Science and Engineering, 2024, 12 (8): 1266. doi: 10.3390/jmse12081266
20	Sharma P, Bishit I, Sur A. Wavelength-based attributed deep neural network for underwater image restoration[J]. ACM Transactions on Multimedia Computing, Communications, and Applications, 2023, 19 (1): 2. doi: 10.1145/3511021
21	Liu Z, Lin Y, Cao Y, et al. Swin transformer: hierarchical vision transformer using shifted windows[C]// Proc. of the IEEE/CVF International Conference on Computer Vision, 2021: 9992.
22	Yang K, Wang S X, Dai J C, et al. WITT: a wireless image transmission transformer for semantic communications[C]// Proc. of the IEEE International Conference on Acoustics, Speech and Signal Processing, 2023.
23	Wang Q L, Wu B G, Zhu P F, et al. ECA-Net: efficient channel attention for deep convolutional neural networks[C]// Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020: 11531.
24	Xie S, Girshick R, Dollar P, et al. Aggregated residual transformations for deep neural networks[C]// Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 1492.
25	Shao X L, Chen H R X, Magson K, et al. Deep learning for multi-label classification of coral conditions in the Indo-Pacific via underwater photogrammetry[J]. Aquatic Conservation, 2024, 34 (9): e4241.
26	Liu R S, Fan X, Zhu M, et al. Real-world underwater enhancement: challenges, benchmarks, and solutions under natural light[J]. IEEE Trans. on Circuits and Systems for Video Technology, 2020, 30 (12): 4861. doi: 10.1109/TCSVT.2019.2963772
27	Sheikh H R, Sabir M F, Boivk A C. A statistical evaluation of recent full reference image quality assessment algorithms[J]. IEEE Trans. on Image Processing, 2006, 15 (11): 3440. doi: 10.1109/TIP.2006.881959
28	Richardson T, Kudekar S. Design of low-density parity check codes for 5G new radio[J]. IEEE Communications Magazine, 2018, 56 (3): 28. doi: 10.1109/MCOM.2018.1700839
29	Bourtsoulatze E, Kurka D B, Gunduz D. Deep joint source-channel coding for wireless image transmission[J]. IEEE Trans. on Cognitive Communications and Networking, 2019, 5 (3): 567. doi: 10.1109/TCCN.2019.2919300
30	Zhang W Y, Zhang H J, Ma H, et al. Predictive and adaptive deep coding for wireless image transmission in semantic communication[J]. IEEE Trans. on Wireless Communications, 2023, 22 (8): 5486. doi: 10.1109/TWC.2023.3234408

Bellhop仿真参数	参数值
发射端和接收端距离/m	600
发射端深度/m	50
接收端深度/m	5
中心频率/kHz	24

对比模型	Micro-F1	Macro-F1	健康F1	受损F1	死亡F1
ST	10.4	17.4	3.1	15.6	42.8
RJSCC	16.1	31.3	3.8	19.2	59.7
DJSCC	23.6	37.1	6.6	23.6	69.1

$f$	Micro-F1	Macro-F1
4	0.563	0.339
8	0.565	0.341
16	0.571	0.344
20	0.561	0.342
32	0.553	0.332

模型	Micro-F1	Macro-F1	健康 F1	受损 F1	死亡 F1
w/o MLFM	0.543	0.320	0.835	0.440	0.331
w/o SCM	0.548	0.329	0.829	0.460	0.350
w/o $ {L_T} $	0.524	0.302	0.817	0.425	0.291
w/MLFM-1	0.547	0.324	0.828	0.445	0.338
w/MLFM-2	0.539	0.318	0.821	0.434	0.321
IUMST	0.571	0.344	0.839	0.472	0.377

[1]	金志刚, 王颖, 梁嘉伟, 程思萌, 李根. 云边协同语义架构下的海洋监测新机制[J]. 系统工程与电子技术, 2026, 48(8): 2850-2858.
[2]	金志刚, 洪叶, 苏毅珊, 羊秋玲. 软件定义云边协同架构下的水下监测新机制[J]. 系统工程与电子技术, 2024, 46(3): 1101-1108.

基于语义传输的智能水下监测机制

Intelligent underwater monitoring mechanism based on semantic transmission

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 13

参考文献 30

相关文章 2

编辑推荐

Metrics

本文评价