Research On Amplitude Preserving Processing Method Based On High-resolution Sparse Radon Transform

Amplitude preservation processing seismic data is the basic process to improve the signal-to-noise ratio and resolution of seismic data,and it is also an important guarantee for later migration imaging.In order to save costs,seismic exploration field data acquisition generally adopts sparse sampling methods,accompanied by missing and bad tracks.In addition,due to the influence of terrain,instruments,wind and grass,etc.,the data collected in the field will inevitably contain noise,which seriously affect imaging accuracy.Radon transform is widely used in seismic data processing because of its flexible transform method,and has achieved good results.The main research content of this paper is how to use Radon transform to attenuate random noise and multiple waves,and briefly introduce its application in seismic trace reconstruction,wave field separation,and velocity analysis.Firstly,it summarizes the basic principle and transformation form of Radon transform.Secondly,according to the characteristic difference between the effective signal and different noises(random noise,multiple waves,etc.)in the Radon domain,noises and effective signal can be to separated and attenuated.Based on the parabolic Radon transform,the sparse constrained condition is introduced,and the L₁-L₂ mixed norm is used to solve the objective function,and the sparse constrained high-resolution Radon transform is obtained,which can well realize the separation of random noise and effective signal;it is proposed to use fast hyperbolic Radon transform based on Chirp-Z transform attenuates multiple waves,and achieves ideal results.The following conclusions are drawn specifically:(1)The resolution of conventional Radon transform is not high,and there are phenomena such as spatial aliasing,truncation effect,end effect,etc.,which cannot meet the requirements of actual seismic data preservation processing;the spatial convolution operator is introduced to obtain high-resolution Radon transform,which improves Radon transform resolution;(2)The Radon transform in frequency domain is faster than that in time domain,which greatly improves the calculation efficiency;(3)Radon transform is an inversion problem.Sparse constraints are used to avoid multiple solutions.In other words,using the L₁ norm to constrain the sparse model function and using the L₂ norm to constrain the error term to obtain a sparse solution,which can well separate random noise form significant waves;(4)The conventional hyperbolic Radon changes are time-variant and cannot be transformed into the frequency domain for calculation.This paper uses the substitution method to change the integration path of hyperbolic Radon transform from hyperbola to parabola,a sparse kernel function is introduced,and solved by Chirp-Z transform and fast Fourier transform,which improves the calculation efficiency and improves the calculation accuracy.The effect of attenuation multiples is ideal and meets the requirements of amplitude preservation processing.