Research On Radon Transform And Its Application In Seismic Data Noise Elimination

For a long time, in order to effectively eliminate noise in seismic data, manytechnical approaches have been considerable development, among of all, the Radontransform which is widely applied and research on noise elimination.Radon transform as the research object of this paper, firstly, the basic theory isintroduced in detail, focusing on the linear Radon transform, parabolic Radontransform and hyperbolic Radon transform, and their calculation forms in the timedomain and frequency domain. Meanwhile, in order to eliminate the spatialconvolution effects of Radon transform, the high-resolution Radon transform of thefrequency domain has been further detailed discussion, and analysis of the amplitudecharacteristics of Radon transform. On this basis, mainly for the coherent noiseelimination in seismic data (including linear noise and multiple noise), introduced thehybrid Radon transform and λ-f domain Radon transform.Hybrid Radon transform used in this paper can separate the noise with the linearevents and the signal with the hyperbolic events at the same time, so that in the hybridRadon domain can simultaneously estimate the effective signal and linear noise, avoidconventional linear Radon domain manual removal. The paper introduced the theory ofhybrid Radon transform, and damped least squares method and high-resolutionmethod were used to transform, conduct comparative analysis of the results of themodel with random noise-free data and random noise data can be seen hybrid Radontransform has unique advantages in terms of linear interference suppression, andhigh-resolution transform has better results than damped least squares method, throughthe application of the real seismic data to verify the effectiveness and practicality of themethod.And multiple noise as another coherent noise, multiple elimination problem hasalways been the focus of the noise elimination study. The parabolic Radon transform iswidely used to multiple noise elimination, the paper improved the conventionalparabolic Radon transform, and introduce λ-f domain Radon transform. Byintroducing the parameter λ, not only improved the computational efficiency, but alsohad better resolution capability for low frequency data. The λ-f domain Radontransform theory is introduced in the paper, while the choice of parameters λ wasdiscussed and introduces the realization process of multiple noise elimination, by comparing the application of model data of constant amplitude and model data withAVO effects, we can see that λ-f domain Radon transform compared to conventionalparabolic Radon transform method has better amplitude characteristics, using themethod of the actual data processing, to deal with the real data by using λ-f domainRadon transform method, proves that this method is effective in the suppression ofmultiple noise. By the application of the real seismic data, the effectiveness of thesetwo methods has been verified.