RANDOM NOISE REDUCTION USING A HYBRID METHOD BASED ON ENSEMBLE EMPIRICAL MODE DECOMPOSITION

We have proposed a novel hybrid random noise reduction method based on ensemble empirical mode decomposition (EEMD) and wavelet threshold filtering. Firstly, the frequency band ranges of effective signal and noise in the initial seismic data set are studied by Fourier spectrum analysis method. Secondly, we make use of EEMD method to obtain the intrinsic mode functions (IMFs) of each original noisy trace. Wavelet threshold filtering is then applied to the high frequency IMFs of each trace to acquire the new denoised high frequency IMFs. Finally, by stacking the filtered high frequency IMFs with the low frequency IMFs and the trend item, we can obtain the ultimately denoised seismic data set. The proposed approach is confirmed via two synthetic data examples and one field data example. The results demonstrate that the proposed approach can achieve much cleaner denoising performance without harming most useful signals.