Elimination Of Coherent Interference In Pre-stack Seismic Signal Based On Contourlet

With the addition of seismic exploration complication, the removal of noise in the pre-stake seismic signal is becoming a hot spot of seismic data processing, and organized noise is one of the most serious noises in impact on the effective signal. So the clean noise-removal seismic data is one of the keys of the good interpretation on seismic data.Firstly, it introduces the formation of seismic signals and the basic classification of organized noises, especially in the characteristics of frequency, speed and two-dimensional imaging. The common processing methods and the advantages, disadvantages and suitable environment of them are also described.Secondly, multi-scale analysis is introduced. By analyzing the advantages and disadvantages of WFT, wavelet transform and the characteristic of multi-scale traneform, Contourlet transformation is expatiated. It explans the implementation and the role of component elements of Contourlet transform in detail. The sparsity of Contourlet is carried out by simulation experiments. It selects the same numbers of coefficients to reconstruct images, in the reference of wavelet transform, the sparse of the Contourlet transformation is verified.In the third chapter, it applies Contourlet to the removal of relative interference. The dominant scales and the directional subbands of the relative interferences are determined on the basis of the speed and frequency. Theoretical models and real seismic signal are both processed on these the dominant directional subbands by common methods. According to the shortcomings of the effection, it uses the effective attenuation of the signal trends in the characteristics of the scale to improve the threshold, the effection is proved.In the end, Contourlet is applied to the removal of surface wave. According to the characteristic of Contourlet transform and the frequency of surface wave, the seismic data is pre-processed. By the effective signal energy ratio of the scale distribution, the fitting function is used to choose to the suited scales and design threshold. Theoretical models and real seismic signal are both processed, and the effect is improved. At the same time, compared with traditional methods and analysis, its superiority is proved.