Tomographic updating for esidual velocities

摘要

The depth mismatch of events at different offsets on a collection of common image gathers (CIGs) is minimized by back-projection of slowness values in a grid. Rather than laboriously picking events on given CIGs, (v,t) pairs are used to calculate necessary arrival times and depths. For each iteration of the prestack migration process, the prvious velocity field is used to calculate ray paths, change in path length, slowness, and opening angles. Back-projection is constrained by 9a) smoothing slowness laterally and vertically to stabilize the minimization, and (b) holding slowness constant in certain regions (e.g. salt bodies). The utility of the approach is shown through application to both finite-difference synthetic and real data in two and threee dimensions.