Reflection waveform inversion enhanced by reverse time demigration-reconstructed observed data

ABSTRACT Reflection waveform inversion (RWI) enables robust background velocity estimation by minimizing the waveform differences between observed and synthetic seismic data. However, synthetic data generated via migration/demigration often deviate unexpectedly from the observed data due to unknown source wavelets and inaccurate reflection coefficients. To mitigate these effects, we use single-shot reverse time migration and demigration operators to reconstruct the observed reflected data before implementing RWI. Due to the conjugacy of the operators, the reconstructed data can preserve kinematic information, even in the case of migration with an inaccurate velocity model. This reconstruction process not only efficiently eliminates the impact of inaccurate reflection coefficients on waveform matching but also guarantees the coherence of the reconstructed data with synthetic data in seismic wavelets. Using this reconstruction process, we use reconstructed data for RWI, enabling us to identify waveform differences accurately. Our approach has been demonstrated successfully using 2D numerical examples.