You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Deepwave currently contains forward modeling of the full wavefield, which is
needed for FWI, but seismic imaging, such as LSRTM, needs forward modeling of
the scattered wavefield. This can be obtained using Born forward modeling. This
would use a wave speed model to propagate a source wave as usual,
creating a wavefield u_0, but would also use a scattering amplitude model,
which, in combination with u_0, would be used to calculate the sources for
the scattered wavefield, u_sc. Backpropagation also needs to be defined
so that the scattering amplitude model can be updated.
The propagator should check whether the scattering amplitude model is all
zeros, in which case the scattered wavefield does not need to be calculated
as it will be zero. This is generally the case for RTM and the first iteration
of LSRTM.
The text was updated successfully, but these errors were encountered:
Deepwave currently contains forward modeling of the full wavefield, which is
needed for FWI, but seismic imaging, such as LSRTM, needs forward modeling of
the scattered wavefield. This can be obtained using Born forward modeling. This
would use a wave speed model to propagate a source wave as usual,
creating a wavefield u_0, but would also use a scattering amplitude model,
which, in combination with u_0, would be used to calculate the sources for
the scattered wavefield, u_sc. Backpropagation also needs to be defined
so that the scattering amplitude model can be updated.
The propagator should check whether the scattering amplitude model is all
zeros, in which case the scattered wavefield does not need to be calculated
as it will be zero. This is generally the case for RTM and the first iteration
of LSRTM.
The text was updated successfully, but these errors were encountered: