| ELECTROMAGNETISM, OPTICS, ACOUSTICS, HEAT TRANSFER, CLASSICAL MECHANICS, AND FLUID DYNAMICS |
Prev
Next
|
|
|
Surface and underwater target classification under limited sample sizes based on sound field elevation structure |
| Yixin Miao(苗艺馨)1,2,3, Jin Fu(付进)4,†, and Xue Wang(王雪)5 |
1 National Key Laboratory of Underwater Acoustic Technology, Harbin Engineering University, Harbin 150001, China;
2 Key Laboratory of Marine Information Acquisition and Security (Harbin Engineering University), Ministry of Industry and Information Technology, Harbin 150001, China;
3 College of Underwater Acoustic Engineering, Harbin Engineering University, Harbin 150001, China;
4 School of Electronic Information Engineering, Beihang University, Beijing 100094, China;
5 Hangzhou Applied Acoustics Research Institute, Hangzhou 310023, China |
|
|
|
|
Abstract Surface/underwater target classification is a key topic in marine information research. However, the complex underwater environment, coupled with the diversity of target types and their variable characteristics, presents significant challenges for classifier design. For shallow-water waveguides with a negative thermocline, a residual neural network (ResNet) model based on the sound field elevation structure is constructed. This model demonstrates robust classification performance even when facing low signal-to-noise ratios and environmental mismatches. Meanwhile, to address the reduced generalization ability caused by limited labeled acoustic data, an improved ResNet model based on unsupervised domain adaptation (“proposed UDA-ResNet”) is further constructed. This model incorporates data on simulated elevation structures of the sound field to augment the training process. Adversarial training is employed to extract domain-invariant features from simulated and trial data. These strategies help reduce the negative impact caused by domain differences. Experimental results demonstrate that the proposed method shows strong surface/underwater target classification ability under limited sample sizes, thus confirming its feasibility and effectiveness.
|
Received: 21 April 2025
Revised: 08 June 2025
Accepted manuscript online: 11 June 2025
|
|
PACS:
|
43.30.-k
|
(Underwater sound)
|
| |
43.30.Bp
|
(Normal mode propagation of sound in water)
|
| |
43.60.Np
|
(Acoustic signal processing techniques for neural nets and learning systems)
|
| |
07.05.Mh
|
(Neural networks, fuzzy logic, artificial intelligence)
|
|
| Fund: This work was supported by the National Natural Science Foundation of China (Grant Nos. 62471024 and 62301183) and the Open Research Fund of Hanjiang Laboratory (KF2024001). |
Corresponding Authors:
Jin Fu
E-mail: fujin@buaa.edu.cn
|
Cite this article:
Yixin Miao(苗艺馨), Jin Fu(付进), and Xue Wang(王雪) Surface and underwater target classification under limited sample sizes based on sound field elevation structure 2025 Chin. Phys. B 34 114301
|
[1] Cheng Y S, Qiu J X, Liu Z and Li H T 2019 Appl. Acoust. 38 653
[2] Wang D Z and Shang E C 2013 Underwater Acoustics 2nd Edn. (Beijing: Science Press) pp. 124–130, 345–347 (in Chinese)
[3] Yang T 2015 J. Acoust. Soc. Am. 138 1678
[4] Li X B, Sun C and Liu X H 2022 Acta. Phys. Sin. 71 233 (in Chinese)
[5] Miao Y X, Fu J, Zou N, Zhao C P and Dong W F 2025 Appl. Acoust. 230 110412
[6] Zheng G Y and Zhu F W 2021 Acoust. Aust. 49 105
[7] Li Z, Gong L J and Li N S 2024 Appl. Ocean. Res. 152 104201
[8] Cao X, Togneri R, Zhang X and Yu Y 2018 IEEE. Sens. J. 14 1
[9] Emmetiere R, Bonne J, Cristol X, Géhant M and Chonavel T 2019 IEEE. J-STSP. 13 185
[10] Conan E, Bonnel J, Nicolas B and Chonavel T 2017 J. Acoust. Soc. Am. 142 2776
[11] Guo X W, Zheng G Y and Yan Q 2022 Appl. Acoust. 198 108985
[12] Cheng Y S, Li Z Z and Qiu J X 2018 Underwater Acoustic Target Recognition (Beijing: Science Press) pp. 12–14 (in Chinese)
[13] Choi J, Choo Y and Lee K 2019 Sensors 19 3492
[14] Luo X W, Qin X M, Wu Z Y, Yang F N, Wang M W and Shang J H 2019 IEEE Access 7 98331
[15] Zhou X Y, Yang K D and Duan R 2019 IEEE. Signal. Proc. Lett. 26 1378
[16] Niu H Q, Gong Z X, Ozanich E, Gerstoft P,Wang H B and Li Z L 2019 J. Acoust. Soc. Am. 146 211
[17] He K M, Zhang X Y, Ren S Q and Sun J 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 26–July 1, 2016, Las Vegas, USA, p. 770
[18] Gao Y J, Chen Y C, Wang F Y and He Y L 2020 IEEE 5th International Conference on Image, Vision and Computing (ICIVC), July 10–12, 2020, Beijing, China, p. 215
[19] Liu F, Ding H, Li D H, Wang T, Luo Z L and Chen L 2021 OES China Ocean Acoustics (COA), July 14–17, 2021, Harbin, China, p. 992
[20] Ganin Y and Lempitsky V 2015 Proceedings of the 32nd International Conference on Machine Learning (ICML), July 6–11, 2015, Lille, France, p. 1180
[21] Porter M B 1991 The KRAKEN Normal Mode Program (La Spezia: SACLANT Undersea Research Center) Technical Report SM-245
[22] Information and data for the SWellEx96 experiment
[23] Lee S and Makris N C 2006 J. Acoust. Soc. Am. 119 336
[24] Jensen F B, Kuperman W A, Porter M B and Schmidt H 2011 Computational Ocean Acoustics (New York: Springer) pp. 338–341
[25] Li S Y, Cheng L, Li J, Wang Z C and Li J L 2024 J. Acoust. Soc. Am. 155 3410
[26] Cheng L, Ji X Y, Zhao H F, Li J L and Xu W 2022 J. Acoust. Soc. Am. 151 269
[27] Tzeng E, Hoffman J, Zhang N, Saenko K and Darrell T 2014 arXiv 1412.3474
[28] Sun B, Feng J and Saenko K 2016 Proceedings of the AAAI conference on artificial intelligence, February 12–17, 2016, Phoenix, Arizona USA, p. 2058
[29] Brekhovskikh L M and Lysanov Y P 1990 Fundamentals of Ocean Acoustics (New York: American Institute of Physics Press/ Springer) pp. 131–137
[30] Lin J 2002 IEEE T. Inform. Theory 37 145 |
| No Suggested Reading articles found! |
|
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
Altmetric
|
|
blogs
Facebook pages
Wikipedia page
Google+ users
|
Online attention
Altmetric calculates a score based on the online attention an article receives. Each coloured thread in the circle represents a different type of online attention. The number in the centre is the Altmetric score. Social media and mainstream news media are the main sources that calculate the score. Reference managers such as Mendeley are also tracked but do not contribute to the score. Older articles often score higher because they have had more time to get noticed. To account for this, Altmetric has included the context data for other articles of a similar age.
View more on Altmetrics
|
|
|
