Full-Waveform Inversion with a Pre-Conditioned Approach of Love-Wave in Near-Surface Prospecting

Summary Compared with Rayleigh-wave, Love-wave Full-waveform inversion (FWI) framework is simpler, but has higher precision and resolution in estimating the S-wave velocity of shallow subsurface. In classic gradient-based FWI algorithm, the calculated gradient operator contains a singular value and uneven energy distribution, both of which are caused by a limited frequency band in seismic records, insufficient wavefield illumination, and double scattering. Hence, model parameters cannot be updated with apparent improvements. In order to smother the artifact in gradient operator caused by double scattering and compensate the illumination capability of wavefield, we developed a preconditioned approach based on the inverse scattering theory for Love-wave FWI. Love-wave noisefree and noise-contaminated data from a complex structure model is tested, the inversion results indicate that the pre-conditioned approach greatly enhances the wavefield illumination in deep medium and improves the imaging accuracy compared with the classic gradient-based algorithm and has strong anti-noise capability.