CEEMD and wavelet transform jointed de-noising method

Abstract

Empirical Mode Decomposition(EMD)is a powerful tool for seismic signal analysis,and is widely used in the field of seismic data denoising.As a modified algorithm of the EMD,Complete Ensemble Empirical Mode Decomposition(CEEMD)not only effectively solves the modal aliasing effect in the EMD,but also maintains the EMD advantages for processing nonstationary signal,such as adaptivity,binary filtering characteristic,orthogonality and so on.CEEMD can decompose a given signal into a series of Intrinsic Mode Function(IMF)components,and the components meet the series distribution from high frequency to low frequency,random noise is generally contained the first several high-frequency IMFs. The conventional CEEMD denoising method directly abandons the noisy highfrequency IMF components,which can lead to the loss of effective high-frequency signal.On the other hand,wavelet threshold denoising method can remove most of random noise,but inevitably suppress the weak effective signal due to its own characteristic. Considering the non-stationarity of seismic signal and the adaptability of denoising method to nonstationary signal,in this paper we combine CEEMD denosing method with wavelet threshold denosing method,and propose a wavelet threshold denoising method based on CEEMD to make up for the disadvantages of the two denoising methods.In our method,we first use the autocorrelation curve method to select the noisy high-frequency IMF components which used to be abandoned for CEEMD denosing method, then wavelet threshold method is used to denoise the noisy high-frequency IMFs,in order to retain the IMFs' high-frequency effective information,which will be used to reconstruct the original signal with the non-noisy low-frequency IMFs.Applying our proposed method to synthetical and real field data sets indicates that it can significantly improve the signal-to-noise ratio,meanwhile,also greatly maintain the effective signal.Wavelet threshold denoising method based on CEEMD is proved to be an effective and relatively preserved-amplitude denoising way.