Surface‐related multiple attenuation by prediction‐muting parabolic Radon transform

Abstract

Multiple is a coherent noise and exists in almost all seismic data. Based on the principle of parabolic Radon transform (PRT) and wave‐equation multiple prediction, the prediction‐muting parabolic Radon surface‐related multiple attenuation(PMSRMA) approach is proposed in this paper. PMSRMA is prestack seismic data preprocessing for multiple elimination, and also, it combines the advantages of PRT with wave equation based surface‐related multiple attenuation (SRMA) approach. The wavelet and subsurface velocity is not needed, and also no human muting is needed as in conventional PRT scheme. All order multiples can be predicted from the square of the initial shot records in frequency domain, and the travel time of multiple is accurately acquired. Designing two dimensional filter function reasonably in Radon image domain, the automatic multiple muting processing is illustrated in this paper, and also, Levinson recursive algorithm is applied to solution of equation in frequency domain. The multiple attenuation results of theoretical homogeneous layered media model and the complex SMAART shot records demonstrate that PMSRMA approach is high efficiency, high precision and applicable.