水下目标信号的Lofar谱图特征的主分量分析研究

Principal Component Analysis on Lofar Spectrum Features of Underwater Target Signals

研究一种基于 Lofar谱图特征和主分量分析的水下目标信号的特征处理方法。首先介绍了这种方法的应用背景 ,给出了舰船辐射噪声的结构和 Lofar谱特征的提取方法。在简单介绍了主分量分析的有关基础知识和数学解法后 ,对于海上测量获得的舰船辐射噪声信号进行特征提取 ,利用主分量分析的方法对提取的特征向量进行降维处理。针对主分量分析处理前后的特征向量 ,采用结构自适应模糊聚类神经网络分类器分类 ,与直接对Lofar谱特征分类相比 。

A method for processing the features of underwater target signal based on Lofar spectrum feature extraction and principal component analysis is discussed. Features of Lofar spectrum from sonar signals are extracted and analyzed. Major theories and solutions about PCA are introduced and signals of ship-rediated noise are processed based on it. Feature dimensions of Lofar spectrum are successfully reduced and a result for the feature expression is obtained. The data are also applied to fuzzy integrated neural network classifier and it is successful than the result without processing of PCA.