为获取能够增强不同类型舰船区分能力的特征，本文提出一种针对单通道舰船辐射噪声特征的多目标优化提取方法。首先，采用集合经验模态分解（Ensemble Empirical Mode Decomposition，EEMD）方法分解单通道舰船辐射噪声（Ship Radiated Noise，SRN）信号，获取各阶本征模态函数（Intrinsic Mode Function，IMF）分量；然后，从IMF分量中提取频率及能量特征，组建单通道SRN特征向量，并将其视为具备多维特征的最小化多目标优化函数；最后，依据主成分分析（Principal Component Analysis，PCA）方法，通过计算各特征向量的贡献率完成特征向量优化提取，并将提取结果以热图的方式呈现。实验结果显示，该方法可成功获取按频率从高到低排列的IMF分量，并深入分析其中频率特征和能量特征，实现对特征的优化提取。
To obtain features that enhance the ability to distinguish different types of ships, this study proposes a multi-objective optimization extraction method for single channel ship radiated noise features. Firstly, the ensemble empirical mode decomposition (EEMD) method is used to decompose the single channel ship radiated noise (SRN) signal and obtain the intrinsic mode function (IMF) components of each order; Then, frequency and energy features are extracted from the IMF components to construct a single channel SRN feature vector, which is regarded as a minimum multi-objective optimization function with multidimensional features; Finally, based on the principal component analysis (PCA) method, feature vector optimization extraction is completed by calculating the contribution rate of each feature vector, and the extraction results are presented in the form of a heatmap. The experimental results show that this method can successfully obtain IMF components arranged from high to low frequency, and deeply analyze the frequency and energy features to achieve optimized feature extraction.
2025,47(3): 158-162 收稿日期：2024-10-31
DOI：10.3404/j.issn.1672-7649.2025.03.026
分类号：TB566
基金项目：2022年重庆市教育委员会人文社会科学研究项目（ 22SKGH493）
作者简介：孙宝刚(1979-),男,副教授,研究方向为数据挖掘与分析、深度学习
参考文献：
[1] 吴承希, 王彪, 徐千驰, 等. 基于小波包分解和PCA-Attention-LSTM的舰船辐射噪声识别技术[J]. 声学技术, 2022, 41(2): 264?273.
[2] 葛轶洲, 姚泽, 张歆, 等. 水声目标的MFCC特征提取与分类识别[J]. 计算机仿真, 2024, 41(2): 13?16+33.
[3] 郭政, 赵梅, 胡长青, 等. 舰船辐射噪声广义多尺度数学形态学特征提取与应用研究[J]. 振动与冲击, 2022, 41(4): 21?28+100.
[4] 刘丹, 赵梅, 胡长青. 基于VMD与共振稀疏分解的舰船辐射噪声窄带特征提取[J]. 声学技术, 2024, 43(2): 172?181.
[5] ENVELOPE G L P , HOU Y , ENVELOPE H Y P . A novel method for frequency feature extraction of ship radiated noise based on variational mode decomposition, double coupled Duffing chaotic oscillator and multivariate multiscale dispersion entropy[J]. Alexandria Engineering Journal, 2022, 61(8): 6329?6347.
[6] 沈鑫玉, 陈涛, 郭良浩, 等. 遗传算法优化变分模态分解提取舰船辐射噪声特征线谱方法[J]. 应用声学, 2024, 43(1): 1?11.
[7] 王朝, 黄慧涛, 张晶, 等. 基于解空间降维的大规模约束多目标进化算法[J]. 电子学报, 2023, 51(11): 3120?3127.
[8] 李紫鹏, 纪永强, 郭兵勇, 等. 基于特征模式分解的水声目标特征提取方法[J]. 哈尔滨工程大学学报, 2023, 44(9): 1542?1548.
