基于小波变换和神经网络的舰船目标识别和分类  

Recognition and Classification Methods of Ship Targets Based on Wavelet Transform and Neural Network

作　　者：马育锋[1,2] 龚沈光[2] 张凡[2] 李松[2] 

机构地区：[1]湖北省环境保护厅,武汉430072 [2]海军工程大学兵器工程系,武汉430033

出　　处：《系统仿真学报》2009年第23期7598-7600,共3页Journal of System Simulation

摘　　要：根据舰船轴频电场信号的特点,提出了基于小波变换和神经网络的目标识别和分类方法。利用小波变换计算轴频电场信号的功率谱,提取特征量,根据轴频电场平均累积功率谱对舰船目标和海洋环境进行识别。根据轴频电场信号的功率谱分析,对不同类型舰船目标进行识别。运用BP神经网络,对不同类型舰船目标和海洋环境进行分类。通过对海上试验采集轴频电场数据的仿真计算,证实了该识别和分类方法的有效性。According to the characteristic of shaft-rate electric field of ships,the recognition and classification methods of ship targets based on wavelet transform and neural network were proposed.Power spectrum of shaft-rate electric field was computed using wavelet transform,characteristic variable was extracted and targets from sea environment was distinguished based on average cumulate power spectrum of shaft-rate electric field of ships.The different kinds of ship targets could be distinguished by the analysis of power spectrum of shaft-rate electric field of ships.The different kinds of ships and sea environment were classified by BP neural network.The acquisition data of the sea test was simulated.The results show that the methods of recognition and classification are effective.

关 键 词：小波变换 神经网络 轴频电场 目标识别 目标分类 

分 类 号：TN911.7[电子电信—通信与信息系统]