Font Size: a A A

Denoising Research Of Seismic Data Based On EMD-ICA Method

Posted on:2018-04-11Degree:MasterType:Thesis
Country:ChinaCandidate:C X LiuFull Text:PDF
GTID:2370330596468452Subject:Geological Resources and Geological Engineering
Abstract/Summary:PDF Full Text Request
In seismic exploration,there are usually random noise in seismic signal due to a combination of varies of unpredictable factors.And how much the random noise is could affect the signal-to-noise ratio of the signal.In a result,how to increase the SNR of seismic data is always the goal people have been pursuing for too much noise resist in data could reduce the reliability and limit the parameter being extracted.So this paper mainly studied the method of how to eliminate the random noise in seismic data.That Empirical Mode Decomposition(EMD)method could break down the signal into a series of different scales of intrinsic mode function(IMF)signal according to the characteristics of the signal itself is an adaptive time-frequency analysis method suiting for analysis of non-stationary signal processing.In the same time,Independent Component Analysis(ICA),which could decompose the signal into independent components with that corresponding actions are took to reduce the noise through optimization algorithm in the lack of a priori information,is a new multidimensional signal processing method developed from theory of blind source separation.The fundamental principle,process of screening and iteration,nature and problems existing of EMD is being elaborated in the second chapter of this paper.Aiming at the existence of the modal aliasing effect in EMD,the principle and realization process of improved algorithms-Ensemble Empirical Mode Decomposition(EEMD)and Complete Empirical Mode Decomposition,which could improve the modal aliasing problems and provide accurate original signal reconstruction and relatively high computational efficiency,is introduced.This paper expounds the mathematical foundation,basic theory and pretreatment method of the ICA algorithm,and studies the principle and realization process in the field of signal separation of ICA algorithm.On this basis,the principle and characteristics of FastICA algorithm are studied to verify its effectiveness for signal separation.After expounding the principle of utilize of CEEMD method and FastICA algorithm in signal-to-noise separation in seismic data respectively,this paper verifies the advantages and disadvantages of for noise suppression.Considering the advantages and disadvantages,the improved EMD-ICA method is put forward to seismic signal de-noising.Simulation experiment confirms that the proposed algorithm could obtain more satisfactory result than independently CEEMD or FastICA or traditional noise reduction method,so as to the feasibility in the application of actual data.
Keywords/Search Tags:Empirical Mode Decomposition, Independent Component Analysis, random noise, signal-to-noise ratio, de-noising
PDF Full Text Request
Related items