Underwater Acoustic Target Recognition Based on Deep Residual Attention Convolutional Neural Network
<p>Time domain waveform and Spectrogram of underwater acoustic target radiated noise. (<b>a</b>) Time-domain waveform of target I; (<b>b</b>) Time-frequency spectrum of target I; (<b>c</b>) Time-domain waveform of target II; (<b>d</b>) Time-frequency spectrum of target II; (<b>e</b>) Time-domain waveform of target III; (<b>f</b>) Time-frequency spectrum of target III.</p> "> Figure 1 Cont.
<p>Time domain waveform and Spectrogram of underwater acoustic target radiated noise. (<b>a</b>) Time-domain waveform of target I; (<b>b</b>) Time-frequency spectrum of target I; (<b>c</b>) Time-domain waveform of target II; (<b>d</b>) Time-frequency spectrum of target II; (<b>e</b>) Time-domain waveform of target III; (<b>f</b>) Time-frequency spectrum of target III.</p> "> Figure 2
<p>Schematic diagram of underwater acoustic target signal framing.</p> "> Figure 3
<p>The flow of hydroacoustic target recognition method.</p> "> Figure 4
<p>Residual attention convolution block structure.</p> "> Figure 5
<p>Diagram of channel attention model.</p> "> Figure 6
<p>Structure of the DRACNN model.</p> "> Figure 7
<p>Training and validation curves of the DRACNN model. (<b>a</b>) Loss. (<b>b</b>) <span class="html-italic">Accuracy</span>.</p> "> Figure 8
<p>Confusion matrix for recognition results on ShipsEar dataset.</p> "> Figure 9
<p>Recognition <span class="html-italic">accuracy</span> at different signal-to-noise ratios.</p> "> Figure 10
<p>Downscaling and visualization of data using TSNE. (<b>A</b>) Raw signals of ShipsEar dataset. (<b>B</b>) Features of the ShipsEar dataset extracted by GAP layer. (<b>C</b>) Raw signals of DeepShip dataset. (<b>D</b>) Features of the DeepShip dataset extracted by GAP layer.</p> "> Figure 11
<p>A new model whose FEM is untrainable.</p> "> Figure 12
<p>Confusion matrix for recognition results on DeepShip dataset.</p> ">
Abstract
:1. Introduction
2. Under Water Acoustic Target Characteristics and Signal Preprocessing
2.1. Target Characteristics
2.2. Signal Preprocessing
3. UATR Method
3.1. Residual Attention Convolution Blocks
3.2. DRACNN Model
4. UATR Experiment and Analysis
4.1. Experimental Database
4.2. Introduction to the Sample Set
4.3. Experimental Results and Analysis
4.4. Generalization Ability of DRACNN Model
5. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
Abbreviation
Abbreviation | Full name |
UATR | Underwater acoustic target recognition |
DRACNN | Deep residual attention convolutional neural network |
DCGAN | Deep conditional generative adversarial network |
LSTM | Long short-term memory |
RACM | Residual attention convolution module |
SE | Squeeze excitation |
FEM | Feature extraction model |
DCB | Deep convolutional block |
MCC | Multi-convolutional layer jumper connection |
CM | Classification module |
GAP | Global average pooling |
References
- Xu, Y.C.; Cai, Z.M.; Kong, X.P. Improved pitch shifting data augmentation for ship-radiated noise classification. Appl. Acoust. 2023, 221, 109468. [Google Scholar] [CrossRef]
- Li, G.H.; Bu, W.J.; Yang, H. Research on noise reduction method for ship radiate noise based on secondary decomposition. Ocean. Eng. 2023, 268, 113412. [Google Scholar] [CrossRef]
- Esmaiel, H.; Xie, D.; Qasem, Z.A.; Sun, H.; Qi, J.; Wang, J. Multi-Stage Feature Extraction and Classification for Ship-Radiated Noise. Sensors 2021, 22, 112. [Google Scholar] [CrossRef] [PubMed]
- Ni, J.S.; Zhao, M.; Hu, C.Q.; Lv, G.T.; Guo, Z. Ship Shaft Frequency Extraction Based on Improved Stacked Sparse Denoising Auto-Encoder Network. Appl. Sci. 2022, 12, 9076. [Google Scholar] [CrossRef]
- Li, Y.X.; Tang, B.Z.; Jiao, S.B. Optimized Ship-Radiated Noise Feature Extraction Approaches Based on CEEMDAN and Slope Entropy. Entropy 2022, 24, 1265. [Google Scholar] [CrossRef]
- Santos-Domínguez, D.; Torres-Guijarro, S.; Cardenal-López, A.; Pena-Gimenez, A. ShipsEar: An underwater vessel noise database. Appl. Acoust. 2016, 113, 64–69. [Google Scholar] [CrossRef]
- Irfan, M.; Zheng, J.B.; Ali, S.; Iqbal, M.; Masood, Z. DeepShip: An underwater acoustic benchmark dataset and a separable convolution based autoencoder for classification. Expert Syst. Appl. 2021, 183, 115270. [Google Scholar] [CrossRef]
- Chen, J.; Han, B.; Ma, X.F.; Zhang, J. Underwater Target Recognition Based on Multi-Decision LOFAR Spectrum Enhancement: A Deep-Learning Approach. Future Internet 2021, 13, 265–285. [Google Scholar] [CrossRef]
- Hong, G.; Suh, D. Mel Spectrogram-based advanced deep temporal clustering model with unsupervised data for fault diagnosis. Expert Syst. Appl. 2023, 217, 119511. [Google Scholar] [CrossRef]
- Meng, L.X.; Xu, X.L.; Zuo, Y.B. Fault feature extraction of logarithmic time-frequency ridge order spectrum of planetary gearbox under time-varying conditions. J. Vib. Shock. 2020, 39, 163–169. [Google Scholar] [CrossRef]
- Wen, L.; Li, X.; Li, X.; Gao, L. A New Transfer Learning Based on VGG-19 Network for Fault Diagnosis. In Proceedings of the 2019 IEEE 23rd International Conference on Computer Supported Cooperative Work in Design (CSCWD), Porto, Portugal, 6–8 May 2019; pp. 205–209. [Google Scholar] [CrossRef]
- Triyadi, A.B.; Bustamam, A.; Anki, P. Deep Learning in Image Classification using VGG-19 and Residual Networks for Cataract Detection. In Proceedings of the 2022 2nd International Conference on Information Technology and Education (ICIT&E), Malang, Indonesia, 22 January 2022; pp. 293–297. [Google Scholar] [CrossRef]
- Hong, F.; Liu, C.W.; Guo, L.J.; Chen, F.; Feng, H.H. Underwater Acoustic Target Recognition with a Residual Network and the Optimized Feature Extraction Method. Appl. Sci. 2021, 11, 1442. [Google Scholar] [CrossRef]
- Li, J.; Wang, B.X.; Cui, X.R.; Li, S.B.; Liu, J.H. Underwater Acoustic Target Recognition Based on Attention Residual Network. Entropy 2022, 24, 1657. [Google Scholar] [CrossRef]
- Li, P.; Wu, J.; Wang, Y.X.; Lan, Q.; Xiao, W.B. STM: Spectrogram Transformer Model for Underwater Acoustic Target Recognition. J. Mar. Sci. Eng. 2022, 10, 1428. [Google Scholar] [CrossRef]
- Luo, X.W.; Zhang, M.H.; Liu, T.; Huang, M.; Xu, X.G. An Underwater Acoustic Target Recognition Method Based on Spectrograms with Different Resolutions. J. Mar. Sci. Eng. 2021, 9, 1246–1265. [Google Scholar] [CrossRef]
- Gao, Y.; Chen, Y.; Wang, F.; He, Y. Recognition Method for Underwater Acoustic Target Based on DCGAN and DenseNet. In Proceedings of the 2020 IEEE 5th International Conference on Image, Vision and Computing (ICIVC), Beijing, China, 10–12 July 2020; pp. 215–221. [Google Scholar] [CrossRef]
- Hu, G.; Wang, K.J.; Liu, L.L. Underwater Acoustic Target Recognition Based on Depthwise Separable Convolution Neural Networks. Sensors 2021, 21, 1429–1448. [Google Scholar] [CrossRef] [PubMed]
- Li, J.H.; Yang, H.H. The underwater acoustic target timbre perception and recognition based on the auditory inspired deep convolutional neural network. Appl. Acoust. 2021, 182, 108210. [Google Scholar] [CrossRef]
- Song, X.P.; Cheng, J.S.; Gao, Y. A New Deep Learning Method for Underwater Target Recognition Based on One-Dimensional Time-Domain Signals. In Proceedings of the 2021 OES China Ocean Acoustics (COA), Harbin, China, 14–17 July 2021; pp. 1048–1051. [Google Scholar] [CrossRef]
- Yang, H.H.; Li, J.H.; Sheng, M.P. Underwater acoustic target multi-attribute correlation perception method based on deep learning. Appl. Acoust. 2022, 190, 108644. [Google Scholar] [CrossRef]
- Ni, J.S.; Hu, C.Q.; Zhao, M. Recognition method of ship radiated noise based on VMD and improved CNN. J. Vib. Shock. 2023, 42, 74–82. [Google Scholar] [CrossRef]
- Yin, F.; Li, C.; Wang, H.B.; Nie, L.X.; Zhang, Y.L.; Liu, C.R.; Yang, F. Weak Underwater Acoustic Target Detection and Enhancement with BM-SEED Algorithm. J. Mar. Sci. Eng. 2023, 11, 357–373. [Google Scholar] [CrossRef]
- Yao, Q.H.; Wang, Y.; Yang, Y.Y. Underwater Acoustic Target Recognition Based on Data Augmentation and Residual CNN. Electronics 2023, 12, 1206–1222. [Google Scholar] [CrossRef]
- Woo, S.H.; Park, J.; Lee, J.Y.; Kweon, I.S. CBAM: Convolutional Block Attention Module. In Proceedings of the 15th European Conference on Computer Vision (ECCV), Munich, Germany, 8–14 September 2018; Volume 11211, pp. 3–19. [Google Scholar] [CrossRef]
- Malla, P.P.; Sahu, S.; Alotaibi, A.I. Classification of Tumor in Brain MR Images Using Deep Convolutional Neural Network and Global Average Pooling. Processes 2023, 11, 679–695. [Google Scholar] [CrossRef]
- Huang, G.; Liu, Z.; Maaten, L.V.D.; Kilian, Q.W. Densely Connected Convolutional Networks. In Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA, 21–26 July 2017; pp. 2261–2269. [Google Scholar] [CrossRef]
- Pathak, D.; Raju, U. Shuffled-Xception-DarkNet-53: A content-based image retrieval model based on deep learning algorithm. Comput. Electr. Eng. 2023, 107, 108647. [Google Scholar] [CrossRef]
- Ding, X.; Zhang, X.; Ma, N.; Han, J.; Ding, G.; Sun, J. RepVGG: Making VGG-style ConvNets Great Again. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Virtual, 19–25 June 2021; pp. 13728–13737. [Google Scholar] [CrossRef]
- Liu, F.; Shen, T.S.; Luo, Z.L.; Zhao, D.; Guo, S.J. Underwater target recognition using convolutional recurrent neural networks with 3-D Mel-spectrogram and data augmentation. Appl. Acoust. 2021, 178, 107989. [Google Scholar] [CrossRef]
- Ke, X.Q.; Yuan, F.; Chen, E. Underwater Acoustic Target Recognition Based on Supervised Feature-Separation Algorithm. Sensors 2018, 18, 4318–4341. [Google Scholar] [CrossRef] [PubMed]
- Feng, S.; Zhu, X.Q. A Transformer-Based Deep Learning Network for Underwater Acoustic Target Recognition. IEEE Geosci. Remote Sens. Lett. 2022, 19, 1505805. [Google Scholar] [CrossRef]
- Hsiao, S.F.; Tsai, B.C. Efficient Computation of Depthwise Separable Convolution in MobileNet Deep Neural Network Models. In Proceedings of the 2021 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW), Penghu, Taiwan, China, 15–17 September 2021; p. 9602973. [Google Scholar] [CrossRef]
Layer (Block) | Channels | Input Shape | Output Shape |
---|---|---|---|
Input Layer | 1 | \ | (None, 4096, 1) |
RACB-1 | 16 | (None, 4096, 1) | (None, 1024, 16) |
RACB-2 | 32 | (None, 1024, 16) | (None, 256, 32) |
DCB-1 | 32 | (None, 4096, 1) | (None, 256, 32) |
RACB-3 | 64 | (None, 256, 32) | (None, 64, 64) |
DCB-2 | 64 | (None, 1024, 16) | (None, 64, 64) |
RACB-4 | 128 | (None, 64, 128) | (None, 16, 128) |
DCB-3 | 128 | (None, 256, 32) | (None, 16, 128) |
GAP Layer | \ | (None, 16, 128) | (None, 128) |
Dense Layer | \ | (None, 128) | (None, m) |
Output Layer | \ | (None, m) | \ |
Total params: 0.26 M | Flops: 5.12 M |
Category | Type of Vessel | Files | Duration |
---|---|---|---|
Class A | Fishing boats, trawlers, mussel boats, tugboats, dredgers | 17 | 1880 |
Class B | Motor boats, pilot boats, sailboats | 19 | 1567 |
Class C | Passenger ferries | 30 | 4276 |
Class D | Ocean liners, ro-ro vessels | 12 | 2460 |
Class E | Background noise recordings | 12 | 1145 |
Category | Training Set | Test Set |
---|---|---|
Class A | 14,708 | 3638 |
Class B | 12,165 | 3114 |
Class C | 33,247 | 8485 |
Class D | 19,278 | 4733 |
Class E | 9034 | 2140 |
Total | 88,432 | 22,110 |
Experiment Times | Accuracy (%) | Value of Loss Function |
---|---|---|
01 | 97.0 | 0.11 |
02 | 96.8 | 0.12 |
03 | 97.2 | 0.10 |
04 | 97.6 | 0.09 |
05 | 96.9 | 0.12 |
06 | 97.1 | 0.11 |
07 | 97.2 | 0.11 |
08 | 97.0 | 0.13 |
09 | 97.4 | 0.10 |
10 | 96.8 | 0.12 |
Average | 97.1 | 0.11 |
Std | 0.24 | 0.01 |
Category | Precision (%) | Recall (%) | F1_Score (%) |
---|---|---|---|
Class A | 97.7 | 96.8 | 97.3 |
Class B | 97.1 | 94.7 | 95.9 |
Class C | 97.3 | 98.2 | 97.8 |
Class D | 97.8 | 98.4 | 98.1 |
Class E | 99.4 | 99.1 | 99.2 |
Average | 97.9 | 97.5 | 97.7 |
No. | Model | Accuracy (%) | Params (M) | Flops (G) |
---|---|---|---|---|
1 | DenseNet-121 [27] | 90.1 | 6.96 | 0.610 |
2 | DarkNet-53 [28] | 96.6 | 40.59 | 1.930 |
3 | RepVGG-A0 [29] | 97.0 | 7.83 | 0.420 |
4 | CRNN-9 [30] | 91.4 | 3.88 | 0.110 |
5 | Autoencoder [31] | 93.3 | 0.18 | 0.410 |
6 | ResNet-18 [13] | 94.9 | 0.33 | 0.110 |
7 | AResNet [14] | 98.0 | 9.47 | 1.460 |
8 | UATR-Transformer [32] | 96.9 | 2.55 | 0.230 |
9 | MobileNet-V2 [33] | 94.0 | 2.23 | 0.140 |
10 | Our | 97.1 | 0.26 | 0.005 |
Category | Training Set | Test Set |
---|---|---|
Cargo | 9277 | 2363 |
Passenger ship | 8872 | 2240 |
Tanker | 8626 | 2144 |
Tug | 9354 | 2286 |
Total | 36,129 | 9033 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Ji, F.; Ni, J.; Li, G.; Liu, L.; Wang, Y. Underwater Acoustic Target Recognition Based on Deep Residual Attention Convolutional Neural Network. J. Mar. Sci. Eng. 2023, 11, 1626. https://doi.org/10.3390/jmse11081626
Ji F, Ni J, Li G, Liu L, Wang Y. Underwater Acoustic Target Recognition Based on Deep Residual Attention Convolutional Neural Network. Journal of Marine Science and Engineering. 2023; 11(8):1626. https://doi.org/10.3390/jmse11081626
Chicago/Turabian StyleJi, Fang, Junshuai Ni, Guonan Li, Liming Liu, and Yuyang Wang. 2023. "Underwater Acoustic Target Recognition Based on Deep Residual Attention Convolutional Neural Network" Journal of Marine Science and Engineering 11, no. 8: 1626. https://doi.org/10.3390/jmse11081626