An improved approach for suppressing noise based on empirical mode decomposition

Noise attenuation always played very important role in seismic signal processing. The predictive filter designed in temporal frequency (f) and space (x) domain can remove only random noise. This paper presented an improved filtering method for suppressing coherent noise and stochastic disturbance in seismic signals. For every frequency slice in f-x domain, we convert the data into a series of intrinsic mode functions (IMFs) by empirical mode decomposition algorithm. In order to eliminate noise, we remove the first IMF. The method can remove both stochastic disturbance and regular noise with steep dip. Applying in prestack as well as poststack sections compares well with the f-x predictive filtering. Real data examples are provided to show the effectiveness of presented approach.