Full-waveform inversion (FWI) is a challenging data-fitting procedure to extract quantitative information of the underground structure from the full information content in the seismogram. Traditional FWI methods have 2 drawbacks. First, it is sensitive to the initial values for lack of low frequency wave components. Second, it requires a huge amount of computation.
Facing the 2 challenges, we propose a new FWI scheme. By introducing the source encoding method and the ensemble Kalman filter (EnKF), we approximate the wave velocity field gradually along with the nonlinear evolution of its error covariance, leading to larger convergence domain with acceptable amount of computation. The project is in development, and exciting results will be shown online soon.