The Study On Seismic Data Denoising And Interpolation With Curvelet Thresholding Iterative Method

With increasing levels of domestic and international exploration, the exploration continues to expand and exploration targets are becoming complex. Complex geological condition,natural conditions and restrictions of economic conditions and other factors all can affect the results of the exploration. Those may cause some seismic signal interference and distortion, allow the acquisition of seismic data to a more complex type of noise, lower signal to noise ratio. In some cases, the result may also be obtained in the absence of obvious seismic data channel, appear the error in the prediction of multiple wave and the aliasing of image etc. These cases seriously affect the processing and interpretation of seismic data work, reduce the seismic data processing fidelity and resolution. In order to find possible new reserves and improve the efficiency and accuracy of seismic erploration, scholars continue to explore proposed new seismic data processing method to deal with increasingly complex exploration conditions. Denoising and interpolation of seismic data is an important part of the seismic data processing.Curvelet transform solve the feature of high dimension and curve of singular more accurately. The threshold iterative method based on the optimization inversion of the sparse constraint can achieve better result with the increase of the sparsity of the inversion parameters. It has a more prominent advantages in terms of denoising. We can get a better result when combining them in the process of the seismic data analysis.This paper describes both typical for the conventional denoising method (median filtering and FX deconvolution) for Seismic Data,And then add noise, seismic signals Curvelet domain transformation, the transformation from the noise also Curvelet domain,After subtracting the two, and finally the results were compared with the conventional denoising methods to analyze the results.Which were higher signal to noise ratio and signal to noise ratio of low seismic signal denoising Curvelet transform the above were obtained after subtracting the results were compared.On real seismic data for the same treatment, but a step to confirm that the noise of seismic data and transform Curvelet respectively, and finally subtracting the two Curvelet coefficient, the results confirmed Curvelet more step to transform the noise is more obvious in the removal of, with a strong advanta.The computational efficiency of conventional denoising method combines Curvelet is higher than Threshold iteration.We can apply this method to the interpolation of the missing seismic record,meanwhile proceed the process of interpolation by using the Curvelet threshold iterative method again after NMO correction. Compared with the result of the interpolation without NMO correction. The paper first random deletions were25%,50%,75%of the seismic data correction process moving. The numerical computation on the model shows that the Curvelet threshold iterative method can get a better effect on interpolation and its result are more approximate to the initial model as well after the process of NMO correction on the missing seismic trace data of the interpolation.