Research On Seismic Denoising Algorithm Based On The Double Density Dual Tree Complex Wavelet

Seismic signal denoising with extracting useful information is one of the important links in seismic signal processing.The effect of noise suppression exerts an influence on the geological interpretation in the follow-up work.The statistical model is established to achieve random noise suppression of seismic signal in double density dual-tree complex wavelet domain because the double density dual-tree complex wavelet has the properties of good directional selectivity and shift-invariance.The main work is as follows:1.We overview the research background of the seismic signal denoising,significance and research status,and introduce the types of noise characteristics and object evaluation criteria at this stage.2.Briefly Introduce the related theory of discrete wavelet transform;and in order to the defects of discrete wavelet transform,other wavelet transform with more direction is described such as double tree complex wavelet and double density dual-tree complex wavelet transform theory.We focus on the double density dual-tree complex wavelet transform because it has the more directional selectivity and approximate shift-invariant property after we research the double tree complex wavelet and double density dual-tree complex wavelet transform.Furthermore,the decomposition process of several kinds of wavelet transform is further explained by the structure of the filter banks.3.The bivariate model of real and imaginary part based on double density dual-tree complex wavelet transform domain is established because the strong correlation between the real part and imaginary part of wavelet coefficient in the same scale and direction.Therefore,the bivariate model for the real and imaginary parts is established in the double density dual-tree complex wavelet domain,and then the denoised seismic signal is obtained by the maximum posteriori probability estimation.In order to verify the effectiveness of the algorithm described in this article,the synthetic seismograms and actual earthquake records are respectively applied to test,which indicates that this method can effectively preserve the original signal,suppressing random noise,improving the signal noise ratio and resolution.