A simple and efficient inversion approach for estimation of primaries, surface-related multiples, and wavelets

Abstract

Multiples can strongly interfere with primaries if they are not properly handled. An estimation of primaries by sparse inversion (EPSI) method was previously proposed to overcome the inaccuracies in prediction-subtraction techniques. However, the application of the iterative inversion method for the estimation of primaries is limited by the convergence rate and algorithmic complexity, given the expensive computational cost per iteration. A robust EPSI (REPSI) method was recently proposed to eliminate the free parameters involved in the original EPSI method at the expense of employing a complicated spectral-projected gradient l1 (SPGl1) algorithm. In addition, wavelet variation between shots is seldom considered in multiple estimation. Using the feedback primary-multiple model, we propose a simple and efficient inversion approach to estimate primaries, surface-related multiples, and shot-constant and shot-variant wavelets, without any requirements on the subsurface model and assumptions on the wavelet. We ...