AN EFFICIENT WAVEFIELD SIMULATION AND RECONSTRUCTION METHOD FOR LEAST-SQUARES REVERSE TIME MIGRATION

The least-squares reverse time migration (LSRTM) could be seen as an iterative processing of the reverse time migration (RTM). Therefore, a large storage requirement and accurate wavefield reconstruction scheme are also needed by LSRTM when the cross-correlated imaging condition is adopted. We propose a scheme that can reconstruct the wavefield backward accurately in the time domain. The variable order finite difference is introduced near the boundary region, which could reduce the storage requirement needed for the wavefield reconstruction. When the wavefield storage is greatly reduced, the proposed accurate wavefield reconstruction method and the precision of the reconstructed wavefield are feasible for LSRTM. We compare two wavefield reconstruction strategies based on the acoustic wave equation, then give a theoretical comparison of the data storage and computation consumption for the two strategies. At last, we perform a numerical experiment of LSRTM to prove the feasibility of our scheme.