Full-time domain matching pursuit and empirical mode decomposition based sparse fixed-point seismic inversion

Abstract Sparse seismic inversion is wildly utilized for reservoir prediction and resolution improvement. Matching pursuit (MP) an effective algorithm solving L0-norm obtaining sparse results. based on MP (MPSI) can estimate the parameters of subsurface from observations by controlling iterations threshold. However, low convergence stability limit application MPSI. To accelerate MP, full-time domain matching (FTMP) first proposed. The FTMP realize multi-point simultaneously instead searching results one one, which process Also, prior model constraint then involved in objective function to improve layer-boundary fidelity Furthermore, empirical mode decomposition (EMD) introduced into framework recover features signal noisy signal. fixed-point (FP) adopted solve this study. optimal be estimated after finite FP algorithm. Combining framework, constraint, EMD algorithms, a complete method named pursuit-based ultimately synthetic field examples are demonstrate practicality approach. Compared with MPSI, proposed have higher layer-boundary.