利用方差分析法实现水中目标的判别  

Identifying underwater targets based on analysis of variance

作　　者：赵绪明[1] 杨根源[1] 杨汉民[1] 黄晅[1] 

机构地区：[1]海军航空工程学院,山东烟台264001

出　　处：《电光与控制》2008年第1期39-41,91,共4页Electronics Optics & Control

摘　　要：对潜艇目标存在性的判别一直是困扰反潜人员的一个课题,高阶累积量是处理高斯噪声的有效工具,分析海洋环境噪声和潜艇目标噪声的高阶谱特性,分别提取其双谱,结果发现其双谱最大值相差在7个数量级以上,说明双谱是判别目标存在性的有效特征。方差分析的实质是分析数据间的差异情况,从海洋噪声中识别潜艇噪声可以看作是分析某一因素变化对样本数据的影响情况。把有目标存在和无目标存在两种情况的海区噪声双谱特征作为样本,进行方差分析,发现当只有环境噪声时,零检验假设条件成立,而当有潜艇目标存在时,其F值远大于1,零检验假设条件不成立,表明该方法可用于判别目标的存在性。利用Jarque-Bera检验法,可验证样本数据满足总体方差相等的正态分布条件。To distinguish the submarine from marine ambient noise is a subject for the anti-submarine personnel. High-order cumulant is effective for processing Gaussian noise. We analyzed the high-order spectrum characteristics of marine ambient noise and submarine target noise, then picked up their bispectnma. We found that the difference of the bispectnma is larger than 7 magnitude levels. So we can say that bispectnma is an effective characteristic for distinguishing submarine from marine ambient noise. The analysis of variance is to analyze the difference between the data. To distinguish the submarine noise from the marine ambient noise is regarded as to analyze the difference between the two kinds of sample data. There are two kinds of samples, one is marine ambient noise and the other is submarine noise, we carried out the analysis of variance and found that when there was only marine ambient noise, the zero test assumed condition hypothesis stood; When there are two kinds of noise, the F-value is much bigger than 1, and the zero tests assumed condition did not stand. Thus we can know whether or not the submarine exists. We also proved that the sample data are independence mutually, and they meet normal distribution condition whose variance is equal.

关 键 词：水下目标识别 航空反潜 双谱 噪声 方差分析 

分 类 号：V271.4[航空宇航科学与技术—飞行器设计] U674.76[交通运输工程—船舶及航道工程]