基于小波的信噪分离方法

Study on an Approach of Signal Noise Separation Based Wavelet

本文在 Ｄｏｎｏｈｏ 提出的从噪声中恢复信号的非线性小波方法的基本原理的基础上， 并针对非线性小波方法的核心问题——浮动阈值的设定提出了新的算法。通过算例证明， 所设计的浮动阈值在实际应用中效果较好， 特别在低信噪比时信号恢复效果明显优于现常用的浮动阈值设计方法。

In this paper nonlinear wavelet approach to the recovery of signal from noisy data proposed by D.L.Donoho is described. and a new improved algorithm is provided aiming at the key problem of the approach, threshold design. The practical computing examples show that this new algorithm is very efficient, especially when signal noise ratio is lower, This algorithm is evidently better than that often used now.