以提升水下目标识别效果为目的，提出特征提取和最小二乘支持向量机的水下目标识别方法。该方法利用水听器采集水下目标声信号后，通过倒谱方式去除信号内噪声获得水下目标原始信号，再使用Wigner高阶谱方法获得水下目标原始信号的子带能量特征与谱能量特征，并将该特征作输入到最小二乘支持向量机内，利用最小二乘支持向量机分类器获取水下目标识别结果。实验结果表明：该方法可有效去除水下目标声信号内噪声信号，获得波动较为规整的水下目标原始声信号，且提取的水下目标声信号子带能量特征和谱能量特征与其时域较为吻合；识别不同类型水下目标时的精度高达0.978左右，识别水下目标精度较高。
In order to improve the effect of underwater target recognition, an underwater target recognition method based on feature extraction and least square support vector machine is proposed. In this method, the acoustic signal of underwater target is collected by hydrophone, and the original signal of underwater target is obtained by removing the noise in the signal by cepstrum method. Then the sub-band energy characteristics and spectral energy characteristics of the original signal of underwater target are obtained by using Wigner higher-order spectrum method, and the characteristics are input into the least squares support vector machine. The recognition results of underwater targets are obtained by using the least square support vector machine classifier. The experimental results show that this method can effectively remove the noise signal in the acoustic signal of underwater target and obtain the original acoustic signal of underwater target with regular fluctuation. And the sub-band energy characteristics and spectral energy characteristics of underwater target acoustic signal extracted are consistent with their time domain. The accuracy of identifying different types of underwater targets is about 0.978, and the accuracy of identifying underwater targets is high.
2022,44(15): 131-134 收稿日期：2022-03-02
DOI：10.3404/j.issn.1672-7649.2022.15.027
分类号：TN911.7
作者简介：葛召华(1978-)，男，高级工程师，主要从事信息化建设研究
参考文献：
[1] 李琛, 黄兆琼, 徐及, 等. 使用深度学习的多通道水下目标识别[J]. 声学学报, 2020, 45(4): 506–514
LI Chen, HUANG Zhao-qiong, XU Ji, et al. Multi-channel underwater target recognition using deep learning[J]. Acta Acustica, 2020, 45(4): 506–514
[2] 金磊磊, 梁红, 杨长生. 基于卷积神经网络的水下目标声呐图像识别方法[J]. 西北工业大学学报, 2021, 39(2): 285–291
JIN Lei-lei, LIANG Hong, YANG Chang-sheng. Sonar image recognition of underwater target based on convolutional neural network[J]. Journal of Northwestern Polytechnical University, 2021, 39(2): 285–291
[3] 徐建华, 豆毅庚, 郑亚山. 一种基于YOLO-V3算法的水下目标识别跟踪方法[J]. 中国惯性技术学报, 2020, 28(1): 129–133
XU Jian-hua, DOU Yi-geng, ZHENG Ya-shan. Underwater target recognition and tracking method based on YOLO-V3 algorithm[J]. Journal of Chinese Inertial Technology, 2020, 28(1): 129–133
[4] 加尔肯别克, 袁杰. 变样本量学习最小二乘支持向量机算法[J]. 计算机工程, 2019, 45(1): 192–198+205
JIA Er-kenbieke, YUAN Jie. Variable samples learning least square support vector machine algorithm[J]. Computer Engineering, 2019, 45(1): 192–198+205
[5] 张牧行, 申晓红, 何磊, 等. 一种水下目标识别的最大信息系数特征选择方法[J]. 西北工业大学学报, 2020, 38(3): 471–477
ZHANG Mu-hang, SHEN Xiao-hong, HE Lei, et al. Feature selection on maximum information coefficient for underwater target recognition[J]. Journal of Northwestern Polytechnical University, 2020, 38(3): 471–477
[6] 蔡达, 范保杰. 基于空间特征选择的水下目标检测方法[J]. 信息与控制, 2022, 51(2): 214–222
CAI Da, FAN Bao-jie. Spatial feature selection for underwater object detection[J]. Information and Control, 2022, 51(2): 214–222
