Moveout Analysis By Time-warping

The ideal moveout analysis should be dense, azimuthdependent, and nonhyperbolic. Semblance-based methods are an inefficient approach to obtain such a model. We propose an inversion approach to moveout parameter estimation using automatic flattening of common midpoint (CMP) gathers. In contrast to semblance-based moveout analysis, we directly measure local slopes to obtain event geometries, rather than search for the best-fit surfaces through parameter scanning. We use plane-wave destruction filters to automatically measure local slopes throughout a CMP gather. These slopes are used to predict traveltime surfaces, which are then warped until by a non-physical flattening method. By applying the same warping to a time attribute volume, the shifts in time-squared needed to flatten each event can be computed. From these shifts, bestfit physical parameters of any type of moveout curve can be found using a linear or non-linear least-squares fitting scheme.