基于光纤矢量水听器垂直阵列的目标探测识别算法研究  

Research on Target Detection and Identification Algorithm Based on Vertical Fiber Optic Vector Hydrophone Array

作　　者：徐红霞 易朗宇 高发仁 杨廷新 龚海霞 施剑 XU Hongxia;YI Langyu;GAO Faren;YANG Tingxin;GONG Haixia;SHI Jian(The 7th Research Institute of China Electronics Technology Group Corporation,Guangzhou 510310)

机构地区：[1]中国电子科技集团公司第七研究所,广州510310

出　　处：《舰船电子工程》2022年第12期40-44,共5页Ship Electronic Engineering

摘　　要：围绕保护海上平台安全的需求,论文研究了基于光纤矢量水听器垂直阵列的目标探测和目标识别算法,并进行了湖上试验。文中详细介绍了用于目标探测的直方图分析法和用于目标识别的Mel特征与支持向量机。考虑到辐射噪声中不同频率成分蕴含的信息量不同,论文对Mel特征进行了改进,按照负指数规律对不同频率成分进行加权。湖上试验证明,目标探测识别算法的方位估计精度可达到1.5°,目标识别准确率可达90%以上。Surrounding the requirement of protecting the safety of offshore platforms,this paper investigates the target detection and target identification algorithms based on the vertical array of fiber optic vector hydrophones. Besides,on-lake experiments are conducted. The histogram analysis method for target detection together with Mel features and support vector machine for target identification are discussed fully in the paper. Considering the different amount of information carried by different frequency components in the radiated noise,the Mel feature is modified in this paper by weighting the different frequency components according to the negative exponential law. On-lake experiments proves that the target detection and recognition algorithm can achieve an orientation estimation accuracy of 1.5° and a target recognition accuracy of more than 90%.

关 键 词：光纤矢量水听器 目标探测 目标识别 

分 类 号：TN911.7[电子电信—通信与信息系统]