Elastic reverse time migration of Flank-preserving deprimary based on Hilbert transform

Elastic reverse time migration (ERTM) utilizes multi-component seismic data to image not only PP but also PS waves. The incorporation of PS imaging substantially mitigates the multi-solution problem of inversion and interpretation. However, in ERTM, the PP image is affected by backscattering artifacts when the scattering angle is 180° or close to 180°. When the migration velocity model is inaccurate, false images may appear in regions with large velocity gradients in both the PP and PS images. De-primary ERTM (DpERTM) addresses this issue by decomposing the wavefield into up- and down-going wavefields, and subsequently cross-correlates the downgoing-source and the upgoing-receiver wavefields to eliminate false images and low-frequency noise. However, the prismatic and turning waves used for imaging steeply dipping structures are weakened. Based on this analysis, we develop a method of flank-preserving DpERTM (FpdpERTM), which introduces prismatic and turning waves information into the composite wavefield through a specific directional wavefield decomposition method based on the Hilbert transform. Subsequently, we construct the elastic wave dual-composite cross-correlation imaging condition utilizing composite downgoing-source and composite upgoing-receiver wavefields, which significantly enhances imaging of steeply dipping structures. Several numerical examples are evaluated to demonstrate that our method effectively suppresses low-frequency noise and eliminates false images caused by large velocity gradients while preserving steeply dipping structures.