一种基于RBF神经网络预测模型的舰船静电场检测方法  

A Ship′s Static Electric Signature Detection Algorithm Based on the Predictive Model of RBF Neural Network

作　　者：李松[1,2] 蔡旭东[2] 赵仕伟[3] 

机构地区：[1]海军装备研究院 [2]海军工程大学兵器工程系 [3]中国人民解放军61741部队

出　　处：《武汉理工大学学报（交通科学与工程版）》2009年第3期553-556,共4页Journal of Wuhan University of Technology（Transportation Science & Engineering）

基　　金：国防重点实验室基金项目资助(批准号:5144407105JB11)

摘　　要：海洋环境中,舰船静电场信号易受背景噪声的影响,给目标检测带来了困难.通过对静电场的分析,提出了一种低信噪比下检测舰船静电场的方法,首先对信号进行小波分解并在低频段进行重构,然后利用RBF神经网络对海洋环境噪声建立了预测模型,以模型对于接受信号的一步预测误差作为检验统计量来判断目标是否存在.将该方法用于实测舰船静电场信号的检测,取得了较好的效果.In the marine environment, the which brings the difficulty to the det ship＇ eetio s static electric signature is easily influenced by the noise, n. Through the analysis of static electric field of the ship, the paper presents an effectual way for detecting the ship static electric signature in the low SNR. First, the signature is decomposed by wavelet-transform technique and reconstructed at the low-frequency region. Then, a predictive of one step predictive error as the not. The real ship is detected and model is set up by use of RBF neural network. It regards signature testing statistic which is used to judge whether the target exist or practical results show that the method is effective.

关 键 词：舰船静电场 目标检测 小波分解 重构 RBF神经网络 预测模型 

分 类 号：TB559[理学—物理]