未来基于水下无人平台的水声目标探测体系要求平台自身具备目标智能化识别能力，而传统水下目标噪声识别方法需要人工提取泛化能力强的特征数据，且识别过程具有较强的人机交互特性，无法满足这一要求。针对这一问题，本文研究一种基于长短时记忆网络（LSTM）的水下目标噪声智能识别方法，借助深度学习自主学习数据特征的能力，应用长短时记忆网络（LSTM）分别对水下目标噪声的时域时间序列数据、频谱数据、梅尔倒谱（MFCC）数据进行深层次特征提取与识别，并使用实际水声目标噪声信号对该方法进行了验证。结果表明，在上述3种输入数据情况下，采用LSTM长短时记忆模型均能有效实现水下目标噪声特征提取与智能识别。
In the future, the underwater acoustic target detection system based on the unmanned underwater platform requires the platform itself to have the ability of intelligent target recognition, while the traditional method of underwater target noise recognition needs to manually extract the feature data with strong generalization ability. And the recognition process has a strong human-computer interaction characteristics, which can not meet this requirement. To solve this problem, an intelligent underwater target noise recognition method based on long short-term memory network (LSTM) is studied in this paper. The time domain time series data, frequency spectrum data and Mel frequency cepstrum coefficient(MFCC) data of underwater target noise are extracted and recognized by long short-term memory network (LSTM). The method is verified by underwater acoustic target noise signal. The results show that the long-short term memory network adopted in this paper can effectively achieve underwater target noise feature extraction and intelligent recognition under the above three input data conditions.
2019,41(12): 181-185 收稿日期：2018-09-05
DOI：10.3404/j.issn.1672-7649.2019.12.035
分类号：TP391.4
作者简介：张少康(1990-),男,博士研究生,研究方向为海洋环境效应技术
参考文献：
[1] 何国庚. 空泡发生与运动理论及其应用[D]. 南昌: 华中理工大学, 2000.10, 82–86. HE Guo-geng. Theory and application of cavitation formation and motion[D]. Nanchang: Central China University of Technology, 2000.10, 82–86.
[2] 吴国清, 李靖, 陈耀明, 等. 舰船噪声识别(I)—总体框架、线谱分析和提取[J]. 声学学报, 1998, 23(5): 394–400 WU Guo-qing, LI Jing, CHEN Yao-ming, et al. Ship radiated-noise recognition(Ⅱ)—the overall framework, analysis and extraction of line-spectrum[J]. ACTA ACUSTICA, 1998, 23(5): 394–400
[3] 吴国清, 李靖, 陈耀明, 等. 舰船噪声识别(Ⅱ)—线谱稳定性和唯一性[J]. 声学学报, 1999, 24(1): 6–11 WU Guo-qing, LI Jing, CHEN Yao-ming, et al. Ship radiated-noise recognition(Ⅱ)—stability and uniqueness of line spectrum[J]. ACTA ACUSTICA, 1999, 24(1): 6–11
[4] 程玉胜, 王易川, 史广智. 基于现代信号处理的技术的舰船噪声信号DEMON分析[J]. 声学技术, 2006, 25(1): 71–74 CHENG Yu-sheng, WANG Yi-chuan, SHI Guang-zhi. DEMON analysis of underwater target radiation noise based on modern signal processing[J]. Technical Acoustics, 2006, 25(1): 71–74
[5] 李启虎, 李敏杨. 水下目标辐射噪声中单频信号分量的检测:理论分析[J]. 声学学报, 2008, 33(3): 193–196 LI Qi-hu, LI Min-yang. The detection of single frequency component of underwater radiated noise of target: theoretical analysis[J]. Acta Acustica, 2008, 33(3): 193–196
[6] 李启虎. 水下目标辐射噪声中单频信号分量的检测:数值仿真[J]. 声学学报, 2008, 33(4): 289–293 LI Qi-hu. The detection of single frequency component of underwater radiated noise of target: digital simulation[J]. Acta Acustica, 2008, 33(4): 289–293
[7] HAWKES M. Acoustic vector-sensor processing in the presence of a reflecting boundary[J]. IEEE Transactions on signal processing, 2000, 48(11): 2981–2993
[8] HAWKES M. Wideband source localization using a distributed acoustic vector-sensor array[J]. IEEE Transactions on signal processing, 2003, 51(6): 1479–1491
[9] 吴国清, 陈永强. 水声瞬态信号短时谱形态及谱相关法检测[J]. 声学学报, 2000, 25(6): 511–516 WU Guo-qing, CHEN Yong-qiang. The analysis of underwater transient signal and their detection by spectral correlation[J]. ACTA ACUSTICA, 2000, 25(6): 511–516
[10] 彭圆, 申丽然, 李雪耀. 基于双谱的水下目标辐射噪声的特征提取与分类研究[J]. 哈尔滨工程大学学报, 2003, 24(4): 390–394 PENG Yuan, SHEN Li-ran, LI Xue-yao. Bispectrum based feature extraction and classification of radiation noises from underwater targets[J]. Journal of Harbin Engineering University, 2003, 24(4): 390–394
[11] 陈敬军, 陆佶人. 被动声纳线谱检测技术综述[J]. 声学技术, 2004, 23(1): 57–60 CHEN Jing-jun, LU Ji-ren. A review of techniques for detection of line-spectrum in passive sonar[J]. Technical Acoustics, 2004, 23(1): 57–60
[12] 李思纯. 基于矢量水听器的目标特征提取与识别结束研究[D]. 哈尔滨: 哈尔滨工程大学, 2007 LI Si-chun. The feature extraction and targets recognition based on vector hydrophone[D]. Harbin : Harbin Engineering University, 2007
[13] 李新欣. 船舶及鲸类声信号特征提取和分类识别研究[D]. 哈尔滨: 哈尔滨工程大学, 2012 LI Xin-xin. Research on feature extraction and classification of ship noise and whale sound[D]. Harbin : Harbin Engineering University, 2012
[14] 孟庆昕. 海上目标被动识别方法研究[D]. 哈尔滨: 哈尔滨工程大学, 2016 MENG Qing-xin. Research on rassive recognition methods of marine targets[D]. Harbin : Harbin Engineering University, 2016
[15] 杨宏晖, 申昇, 姚晓辉, 等. 用于水声目标特征学习与识别的混合正则化深度置信网络[J]. 西北工业大学学报, 2017, 4, 35(2): 220–225 YANG Hong-hui, SHEN Sheng, YAO Xiao-hui, et al. Underwater acoustic target feature learning and recognition using hybrid regularization deep belief networks[J]. Journal of Northwestern Polytechnical University, 2017, 4, 35(2): 220–225
[16] 陈越超, 徐晓男, 姚晓辉, 等. 基于降噪自编码的水中目标识别方法[J]. 声学与电子工程, 2018, 1: 30–33 CHEN Yue-chao, XU Xiao-nan, YAO Xiao-hui, et al. Underwater target recognition based on denoising autoencoder[J]. Acoustics and Electronic Engineering, 2018, 1: 30–33
[17] 王强, 曾向阳. 深度学习方法及其在水下目标识别中的应用[J]. 声学技术, 2015, 4, 34(2): 138–140 WANG Qiang, ZENG Xiang-yang. Deep learning methods and their applications in underwater targets recognition[J]. Technical Acoustics, 2015, 4, 34(2): 138–140
[18] 张大伟, 章新华, 付留芳等. 基于听觉感知与卷积神经网络识别舰船目标[J]. 声学技术, 2015.12, 34(6):181–184. ZHANG Dawei, ZHANG Xinhua, FU Liufang. Recognition of ships based on auditory sense and convolutional neural networks[J]. 2015.12, 34(6):181–184.
[19] LECUN Y, BENGIO Y, HINTON G. Deep Learning[J]. Nature, 2015, 521(7544): 436–444
[20] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural computation, 1997, 9(8): 1735–1780
[21] GRAVES A. Long short-term memory[M]. Berlin:Springer, 2012:1735-1780.
[22] 舒帆, 屈丹, 等. 采用长短时记忆网络的低资源语音识别方法[J]. 西安交通大学学报, 2017, 10, 51(10): 120–127 SHU Fan, QU Dan. A Speech recognition method using long short-term memory network in low resources[J]. Journal of Xi’an Jiao Tong University, 2017, 10, 51(10): 120–127
[23] 王鑫, 吴际, 等. 基于LSTM循环神经网络的故障时间序列预测[J]. 北京航空航天大学学报, 2017, 10, 51(10): 120–127 WANG Xin, WU Ji. Exploring LSTM based recurrent neural networks for failure time series prediction[J]. Journal of Beijing University of Aeronauties and Astronautics, 2017, 10, 51(10): 120–127
[24] KLAUS Greff, RUPESH K, Srivastava, et al. LSTM: A Search Space Odyssey[C]. IEEE Transactions on Neural Networks and Learning Systems, 2017.10, 28(10), 2222–2232.
[25] ALEX Graves. Learning precise timing with LSTM recurrent networks[J]. Journal of Machine Learning Research, 2003, 3(3): 115–143
[26] IAN Goodfellow, Yoshua Bengio, Aaron Courville.Deep Learning[M]. Journal of Machine Learning Research, vol.3, 379-396.
[27] NITISH Srivastava, GEOFFREY Hinton, et al. Dropout: A Simple Way to Prevent Neural Networks from Overfitting[J]. Journal of Machine Learning Research, 2014(15): 1929–1958
