The Multi-dimensional Complete Ensemble Empirical Mode Decomposition and Its Application in Seismic Denoising

Abstract

Summary Data decomposition has an essential role in seismic data processing and interpretation. So far variety of data decomposition methods have been developed especially based on time frequency analysis. Each of them has its own advantages and disadvantages. But the necessities for a powerful method which successfully overcome the decomposition of non-stationary and none-linear data steel exist. In this paper we use the latest approach in BEMD introduced by Wu and Huang (2009). We updated this method by using the CEEMD and applied it for random noise attenuation from synthetic and real seismic data. MCEEMD is a data decomposition approach which decomposes input data into tow-dimensional IMFs. Each IMF reveals separate features of the data. Multi-dimensional CEEMD can be applied on 2D seismic section or other higher-dimension data analysis. The method is based on a completely different approach from surface fitting methods and by passes major obstacles and difficulties in defining extremes. As an elementary application in seismic data processing the method was used for seismic data denoising. It performed well and significantly attenuated the random noise.