Study On Seismic Data Regularization For Amplitude Preserved Imaging

In practical seismic data acquisition,missing traces,outliers,large interval will appear in the data volume due to the impact of surface conditions,the possible failure of seismic instrument or limit of exploration cost,etc.,so the data acquisited is irregular.Irregular data will produce bad effects to the multiple elimination and migration and at last decrease the follow-up processing and interpolation precision.Seismic data regularization is one of the important steps in the data processing,the methods not only can suppress noise and spatial alias or increase the imaging precision,but also can provide necessary data for reservoir analysis.Seismic data regularization is the reconstruction of the wave field and should keep the amplitude relative relationship in horizontal and vertical direction.What's more,the waveform should be consistent.So the regularization is not simple interpolation in mathematics.The regularization methods for amplitude-preserved seismic imaging should keep the kinematics characteristic of the seismic wave and try the best to preserve the dynamics characteristic such as the amplitude information.Traditional interpolation methods can decrease the interval,suppress the spatial alias and improve the signal noise ratio at the same time.It can be satisfied the basic demands of seismic imaging,but can not meet the demand of amplitude preserved imaging and the reservoir analysis.We studied some methods to reconstruct the missing data,such as near offset interpolation method using multiples,seismic data reconstruction method based on parabolic Radon transform and the reconstruction method based on wave field reconstruction.Near offset reconstruction method using multiples consider the multiples as signal,and pseudo-primary is got from the seismic data which contain multiples using cross correlation method.Then the near offset is reconstructed after using filtering method in frequency domain and root mean square amplitude correction method.The method can reconstruct the missing data using the multiples and increasing the amount of information,so it is different with the methods using the known seismic information.The reconstructed data can be used for seismic imaging effectively.The reconstruction method based on parabolic Radon transform combines with the trace equalization method,so the convergence rate and computation efficiency are improved.The trace equalization takes the AVO characteristics of seismic data into consideration to ensure the reconstruction precision.The method can be expanded to three dimensions and realized 3D seismic data reconstruction.The method is not required accurate velocity and has a certain ability to anti-noise.The seismic data changes very small after parabolic Radon transform and inverse transform,so the method is relative amplitude preserve and can be used as seismic data regularization for seismic imaging.The method based on wave field reconstruction using the information of valid frequency band and iterative method to reconstruct missing data.It obtains the idea of compressed sensing that replace the missing traces with zero and makes the data into a regular grid,then the compute efficiency is enhanced by the fast Fourier transform(FFT)method.Because the FFT is the orthogonal transformation,the two domains which before and after FFT has the same energy,that means the method is amplitude preserved and the reconstructed data not only can meet the requirement for seismic imaging,but also can meet the requirement for reservoir analysis.Through the processing of models and practical data,the seismic data regularization methods can reconstruct the missing data effectively and meet the industry requirement for seismic imaging and reservoir analysis.We discussed the amplitude preserved processing and analyzed the amplitude preservation,anti-noise property and the application range of the three proposed methods.In view of the data regularization purpose and the seismic data characteristic,the choice suggestion of the methods is given.