Hybrid-Encoding-Based Reverse Time Migration of Multiples

Abstract

Multiples propagate along longer paths than primaries and have been incorporated into reverse time migration (RTM) to enhance subsurface illumination. However, RTM of multiples often suffers from severe crosstalk artifacts caused by cross-correlations among unrelated multiple orders. Moreover, the conventional shot-domain RTM framework becomes computationally expensive when a large number of shots are involved. To address these challenges, we propose a hybrid-encoding-based RTM method for multiple imaging. Different multiple orders are first separated using a decomposition technique. Random-phase encoding is then applied to consecutive-order multiples to form supergathers, and linear time-delay encoding is used to compress these supergathers into a limited number of plane-wave gathers. The multiple-based image is subsequently obtained using the RTM operator. Compared with conventional shot-domain RTM of multiples, the proposed plane-wave-domain approach effectively suppresses crosstalk from chaotic-order interference while substantially reducing computational cost. Numerical experiments on the Pluto 1.5 model demonstrate that the method efficiently attenuates coherent crosstalk and improves the imaging of complex subsurface structures.