Application of adaptive step size independent component analysis to seismic signal denoising
Xuemin Tian · Computer Engineering and Applications Journal · 2011
An improved approach based on the modified relative gradient Independent Component Analysis(ICA) is proposed for prestack seismic denoising in this paper.Firstly,this method establishes a nonlinear relationship between the step size function and the measure of signal dependence.This algorithm can adjust the step size function adaptively.On the other hand,it adds a momentum to the separating matrix.The momentum can diminish the effect of the step size function's sharp reduction.Finally,it uses the modified ICA model to reduce prestack seismic random noise.Experimental results demonstrate that the proposed ICA method has good effect on removing random noise in prestack seismic record.In addition,the proposed method can protect the available seismic signal effectively.