Double Plane Wave Least Squares Reverse Time Migration

Images obtained from reverse time migration (RTM) can be improved substantially by employing the least squares migration. This approach, however, is computationally very expensive. Here we propose a new least squares migration using compressed data volumes in coupled plane wave or ray-parameter domain. We call this approach the double plane wave (DPW) least square RTM. Based on the Born modeling operator for shot gathers in the frequency domain, we derive a Born modeling operator for DPW dataset. The DPW Born modeling operator is used in the demigration process for the DPW least squares RTM. The DPW least squares RTM requires much fewer wave field computations than that required by traditional shot gather least squares RTM leading to significant speedup in migration. Additionally, the proposed method utilizes the seismic data in the fully decomposed DPW-frequency domain. As a result, selected parts of the plane wave components and frequency components can be migrated to obtain images. We demonstrate the usefulness of our algorithm using two synthetic datasets.