Robust Q-compensated multidimensional impedance inversion using seislet-domain shaping regularization

Abstract

Acoustic impedance (AI) inversion plays a vital role in seismic interpretation because AI contains valuable information on lithology and contributes to reservoir characterization. However, the effect of anelastic attenuation dissipates the energy and distorts the phase of seismic waves during their propagation in the Earth. Such attenuation-induced effects will degrade the quality of AI inversion unless some preprocessing routines are performed in advance (e.g., inverse Q-filtering). In order to invert for AI from nonstationary seismic data directly and enhance the lateral continuity, we propose a robust Q-compensated multidimensional AI inversion method. We incorporate the Q-filtering operator into the conventional convolution model and solve the inverse problem iteratively, which can avoid some of the errors introduced by those compensation-related processing routines. Furthermore, we incorporate structural information into the inversion processing via seislet-domain nonlinear shaping regularization. Compared with the conventional nonstationary multichannel AI inversion method, our proposed method can accelerate the convergence rate during inversion and further improve lateral continuity and accuracy in the presence of noise. Finally, synthetic and field data are used to validate the effectiveness and robustness of the proposed method. The results demonstrate that the proposed method can retrieve AI from nonstationary seismic data directly with improved efficiency and remove possible artifacts caused by ambient noise.