Seismic Data Reconstruction Approach Based On Parabolic Radon Transform

The purpose of seismic exploration is to get accurate imaging of the subsurface. Due to the impact of environmental and human factors, there is inevitable too large sampling space,data missing or irregular sampling phenomena in the spatial direction of seismic data, which will reduce the quality of pre-stack migration imaging, weaken multiple suppression effect,affect subsequent analysis accuracy. To ensure proper operation of the subsequent seismic data processing, we need to reconstruct the irregular seismic data.Parabolic Radon transform is a robust and effective method for seismic data reconstruction. It can achieve an effective reconstruction result through multiple iterations CMP gathers which is after part of NMO. The method requires less input data, which is simple and easy to implement. In the face of huge amounts of data in pre-stack data processing, the computational efficiency is becoming the key to a good and practical algorithm. When the traditional parabolic Radon transform is used to do seismic data reconstruction, it need the more iterations, too large the calculation, what’s worse, direct matrix inversion algorithm is less efficient in solving the forward parabolic Radon transformation.It will produce aliasing problems and amplitude anomalies during Radon transform because of the traditional curvature parameter sampling criteria.In this paper, we have completed creatively the following aspects. Firstly, we improve the least squares parabolic Radon transform by using the new curvature parameter sampling criteria to suppresses aliasing and amplitude anomalies problem, application Levinson recursive method instead of direct matrix inversion methods to improve the computational efficiency. Secondly, we propose high-resolution parabolic Radon transform seismic data reconstruction method which introduce a sparse constraint matrix to improve seismic data resolution in Radon domain, It can reduce iteration times and computation in seismic data reconstruction. Thirdly, we propose?-f domain weighted parabolic Radon transform method.The new variable ? is introduced to make the Radon transform operator frequency-independent. The transform operator and inverse operator could be computed only once, which can significantly improve the computational efficiency. During the iteration process of ?-f domain parabolic Radon transform, changing weighted coefficients are introduced to improve energy focusing problem in ?-f domain.