基于小波变换的水声信号滤波方法

Underwater Acoustic Signal De-Noising Method Based on Wavelet Transform

为更好地改进低信噪比水声信号的滤波效果,本文针对小波滤波算法展开研究.对小波变换系数使用最大后验概率估计算子进行去噪,达到对原始水声信号滤波的目的,去除大部分的噪声;对含噪信号使用可变阈值函数的小波阈值滤波,进一步提高信噪比.文中给出了两种滤波方法的算法,通过对仿真信号和真实水声信号的滤波实验表明:本文方法比常用的滤波方法均方误差更小,更有利于提取水声信号的特征并进行分类识别.

In order to get a better de-noising performance for underwater acoustic signal,the wavelet de-noising algorithm was studied in the paper.Firstly,an analytical expression of the maximum aposteriori（MAP）was applied to wavelet coefficients for eliminating the original acoustic noise,most of the noise were eliminated in this procedure,and then de-noise the noisy signal with wavelet threshold denoising using a variable threshold function,the de-noising performance was improved again.These two kinds of algorithms are presented in this paper,and the experiment of simulation signal and the real underwater acoustic signal show that the method shown in this paper can get less Mean Square Error than usual de-noising methods,and is favor to the feature extraction and classification identification of the underwater targets.