Reverse Time Migration of Crosswell GPR Data Based on Wavefield Decomposition

Reverse-time migration (RTM) has been widely used for imaging subsurface structure in seismic surveys for hydrocarbon resource explorations. It has been adopted for migrating ground penetrating radar (GPR) data due to the kinematics and dynamics similarity between electromagnetic and seismic waves. Low-frequency migration artifacts happen when the conventional image condition is used, especially when the migration model is not smooth. In this abstract, we decompose the wavefield into left-and right-going wavefield based on the Hilbert transform, and then apply the image condition to generate migration image in which the low-frequency noise are largely eliminated. Synthetic results suggest more clear image can be obtained using the isolated wavefiled to construct RTM image.