一种基于DTW算法的谱图匹配声纳目标识别方法  

Sonar Target Classification Based on Spectrum Matching with DTW Algorithm

作　　者：张宝华 刘虎[2] 高鑫[2] 

机构地区：[1]解放军某部,北京100841 [2]海军潜艇学院航海观通系,山东青岛266071

出　　处：《青岛大学学报（工程技术版）》2010年第3期55-57,共3页Journal of Qingdao University(Engineering & Technology Edition)

摘　　要：为解决由于舰船目标螺旋桨轴频不同造成DEMON谱线谱频率不同而难以直接比对的问题,利用DTW算法将DEMON谱数据根据线谱频率进行对齐,解决了难以比对的问题,并对实测舰船辐射噪声进行了分析。结果表明,利用DTW算法的识别率在原先匹配识别的基础上提高了3%~5%,说明基于DTW算法对谱图匹配声纳目标进行识别是有效的。Propeller shaft frequency affects line spectrum frequency in DEMON spectrum,and so some difficulties are caused in case of difference of propeller shaft frequency for direct comparison of DEMON spectrum.In this paper,DEMON date is aligned by DTW algorithm.This method solves the problem of direct comparison.The analysis of the sea test data shows that,by adopting DTW algorithm,the recognition rate is 3%～5% higher than matching recognition.

关 键 词：DTW DEMON谱 目标识别 

分 类 号：U666.7[交通运输工程—船舶及航道工程]