基于经验模分解的小波阈值滤波方法研究

信号的多分辨经验模分解方法可以解释为以信号极值特征尺度为度量的时空滤波过程。这种时空滤波器充 分保留了信号本身的非线性和非平稳特征,在信号的滤波和去噪中具有较大的优势。本文提出了一种基于经验模分解的小 波阈值滤波去噪方法,并和小波阈值去噪、多尺度EMD滤波效果相比较。实验结果表明了基于经验模分解的小波阈值去 噪具有广泛的适用性和独特的去除非平稳信号的有色噪声的优势。

A Study of Wavelet Threshold Filtering Based on Empirical Mode Decomposition

The multiresolution empirical mode decomposition method of signal can be interpreted as spatial filtering based on the signal's extremum characteristic scale. This spatial filter preserves the nonlinear and nonstationary of signal, and has potential superority in filtering and denoising. The paper proposed a wavelet threshold filtering method based on empirical mode decomposition. By contrast with wavelet threshold filter and multiscale EMD filter, the experimental results clarity the ex- tensive applicability and efficient of this method.