基于正交小波包和K-L变换的水声信号特征提取与识别  

Hydroacoustic Signal Feature Extraction & Recognition Based on Orthogonal Wavelet-Packet and K-L Transform

作　　者：葛永[1] 陈建安[1] 

机构地区：[1]西安电子科技大学,陕西西安710071

出　　处：《情报指挥控制系统与仿真技术》2005年第2期8-11,共4页Information Command Control System and Simulation Technology

摘　　要：在海洋作战环境中,水声目标识别与分类是水声设备与水中兵器系统智能化的关键技术之一,基于小波变换的模式识别方法受到广泛研究与重视,在对水声信号进行正交小波包分析的基础上,以信号在各分解子空间的能量分布构成原始特征空间,运用K-L 变换进行模式识别特征空间的提取,并基于最小欧氏距离设计分类器,提出了一种新的水声信号识别方法,仿真实验验证了算法是有效的。In naval battlefield environment, target recognition is a key technique of underwater acoustical equipment and intelligent weapon. The pattern recognition method based on wavelet is emphasized and studied widely. Through combining orthogonal wavelet packet analyzes and K-L transform in pattern recognition field, a new approach of feature extraction is studied in this paper. Based on minimum Euclidean distance, target identification is distinguished. Simulating & calculating results show the pattern recognition approach is effective.

关 键 词：小波包 能量矢量 特征提取 分类器 

分 类 号：E967.3[军事—军事通信学]