Research On The Technology Of Regularization Of Seismic Wave Reverse-time Migration Data

Seismic exploration includes three aspects, the field data acquisition, data processing and interpretation of seismic data. The quality of the field data acquisition decides whether the next two steps can be executed accurately or not. The process of field data acquisition is as follows:in the preliminary determination of the oil and gas distribution line target area, artificial seismic wave excitation, and field seismic instrument receive seismic signal. But the following three aspects influence collection of regularization data. First of all, due to human and financial constraints, only carry on sparse sampling; secondly, the distribution of survey line is influenced by geographical factors, which are not strictly regular sampling; in addition, due to the unavoidable noise, resulting in some low quality seismic data need to be abandoned. It would encrypt spatial sampling rate, improve the SNR, eliminate the spatial aliasing caused by nonregular sampling, and shorten the production cycle, saving manpower and material resources by rule processing seismic data from practical exploration. seismic section from reverse time migration imaging after seismic data after the rule processing by can reflect the actual underground geological structure bitterly. Therefore, regularization of seismic data processing can provide effective help for oil and gas exploration, has an important significance for actual production.According to the application migration imaging technique before or after the horizontal stacking seismic migration, it can be divided into prestack migration and poststack migration. Prestack migration has more advantages than poststack migration:good image on complex tectonic fracture, avoid the NMO correction stack generated error message, prestack common reflection point gathers can be used for AVO andprestack inversion, and obtain the RMS velocity accurately. This paper use prestack reverse time migration to image the data after rule processing.First of all, this paper analyzes the types of datas needed the rule processing,one is the data which is lack of some traces, the other is regular sampling data needs equidistant interpolation.This paper use radon tracsform to recovery the data which is lack of some traces. On the basis of previous studies, the paper has a detailed study of the mechanism of parabolic radon transform, prove the feasibility of parabolic Radon transform for data reconstruction. Firstly, the data which is lack of some traces need to be do the approximate NMO. Then do several parabolic radon transform and inverse transform, iterate, until the data can meet the requirements.This paper studies three kinds of methods for regular sampling data needs equidistant interpolation:f-x domain trace interpolation, f-k domain trace interpolation and generalized f-k domain trace interpolation(GFKI). GFKI interpolation method is needed to calculated interpolation factor firstly, then zero to the traces needed for interpolation and do Fourier transform to the f-k domain, the raw data is filtered to obtain F-K spectra. Do the inverse Fourier transform to acquire a target data. Contrast the effects from the restore data and the raw data on horizontal layered model with above three methods. GFKI weakened the Spatial aliasing which is easy produced from the f-x domain trace interpolation and f-k domain trace interpolation. Then do the forward simulation with marmousi model and reverse time migration imaging with Sparse data and the reconstruction data after GFKI. The results showed that the latter’s reverse time migration imaging reflected the tectonic characteristics of the model is more clearly and accurately than the former.