摘要
小波多尺度分解是一种有效的信号去噪方法。对于非平稳信号的消噪 ,主要是选取合适的小波及每层小波系数的阈值。笔者讨论了小波消噪的机理 ,并提出一种新的取阈值方法。对于含噪声的非平稳信号 ,信号的慢变化部分主要体现在小波分解的低频系数中 ,而信号的突变部分和噪声主要反映在各层的高频分解系数中 ,利用电子测量中的 3σ准则 ,可有效地将信号突变部分和噪声区分出来 ,从而在消除噪声的同时保留信号的突变部分。
It is a effective method to reduce the noise in signal by wavelet multiscale decomposition. The keys of reducing noise in nonstationary signals are to select good wavelet and thresholds for each layer wavelet decomposition coefficients. In this paper we discuss the principle of reducing noise in nonstationary signal by wavelet analysis and put forward a new method of getting thresholds. For nonstationary signals contained noise, the low frequency wavelet coefficients decomposed represent the parts of low change of signal, the high frequency wavelet coefficients decomposed represent the noise and the saltation of signal. It is very convenient to distinguish the parts of noise and saltation of signal by 3σ rule applied in electronic measurement. As a result, we can reserve the saltation of signal and reduce the noise.
出处
《重庆大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2002年第12期58-61,共4页
Journal of Chongqing University
