一种改进的EMD降噪方法

Improved EMD based de-noising method

提出一种改进的基于经验模态分解(EMD)的降噪方法,将基于EMD阈值降噪方法和Savitzky-Golay滤波降噪方法相结合,改进的方法分别将经过EMD分解后得到的噪声信号的高频模态函数用阈值降噪的方法,低频模态函数用Savitzky-Golay滤波降噪的方法,这样既能保持信号的高频部分,又能较好地保持信号低频部分的光滑特性。仿真实验表明,改进的EMD方法降噪性能要优于单独使用EMD阈值方法或Savitzky-Golay滤波方法。

An improved empirical mode decomposition(EMD) based de-noising method was proposed.The improved method hybridized EMD based threshold de-noising and Savitzky-Golay filtering.Noisy signal was decomposed into intrinsic mode functions(IMFs),then the high frequency IMFs were de-noised by soft threshold method and the low frequency IMFs were de-noised by Savitzky-Golay filtering method.Thus the high frequency part and the smoothing property of the low frequency signal can be retained.The method was tested with simulated data and its de-noising performance is better than that of the EMD based threshold method or the Savitzky-Golay filtering method used alone.