1887

Abstract

Summary

Optimal transport based misfit functions have been introduced to mitigate cycle skipping in full waveform inversion. Recently, we have proposed to compare the discrete graph of the data through optimal transport. This strategy makes possible to interpret non-positive data through optimal transport (OT), while maintaining its convexity property with respect to time and amplitude shifts. Here, we introduce a novel numerical strategy, which makes the computational overcost associated with the graph space OT approach almost negligible compared to a conventional least-squares distance. This is possible through the interpretation of the graph space OT problem as a linear sum assignment problem, for which specific algorithms exist: we select the auction algorithm for its good performance on small scale, dense instances of such problems. We introduce a tuning parameter, based on the estimation of the maximum time shift, to scale the misfit function properly. Application on visco-acoustic synthetic data illustrate the interest of the strategy. From significantly cycle-skipped data, the method retrieves a meaningful estimation of the velocity model, while the conventional least-squares approach and one previously introduced OT based approach both fail. The computational cost increase is only few percents compared to the least-squares misfit function.

Loading

Article metrics loading...

/content/papers/10.3997/2214-4609.201900870
2019-06-03
2024-03-29
Loading full text...

Full text loading...

References

  1. Benamou, J.D. and Brenier, Y.
    [2000] A computational fluid mechanics solution to the Monge-Kantorovich mass transfer problem. Numerische Mathematik, 84(3), 375–393.
    [Google Scholar]
  2. Bertsekas, D.P. and Castanon, D.
    [1989] The auction algorithm for the transportation problem. Annals of Operations Research, 20(1), 67–96.
    [Google Scholar]
  3. Birkhoff, G.
    [1946] Three observations on linear algebra. Univ. Nac. Tucumán, A(5), 147–151.
    [Google Scholar]
  4. Métivier, L., Allain, A., Brossier, R., Mérigot, Q., Oudet, E. and Virieux, J.
    [2018] Optimal transport for mitigating cycle skipping in full waveform inversion: a graph space transform approach. Geophysics, 83(5), R515–R540.
    [Google Scholar]
  5. Métivier, L., Brossier, R., Mérigot, Q. and Oudet, E.
    [2019] A graph space optimal transport distance as a generalization of Lp distances: application to a seismic imaging inverse problem. Inverse Problems, to be submitted.
    [Google Scholar]
  6. Métivier, L., Brossier, R., Mérigot, Q., Oudet, E. and Virieux, J.
    [2016] An optimal transport approach for seismic tomography: Application to 3D full waveform inversion. Inverse Problems, 32(11), 115008.
    [Google Scholar]
  7. Plessix, R.E.
    [2006] A review of the adjoint-state method for computing the gradient of a functional with geophysical applications. Geophysical Journal International, 167(2), 495–503.
    [Google Scholar]
  8. Virieux, J. and Operto, S.
    [2009] An overview of full waveform inversion in exploration geophysics. Geophysics, 74(6), WCC1–WCC26.
    [Google Scholar]
  9. Yang, P., Brossier, R., Métivier, L., Virieux, J. and Zhou, W.
    [2018] A Time-Domain Preconditioned Truncated Newton Approach to Multiparameter Visco-acoustic Full Waveform Inversion. SIAM Journal on Scientific Computing, 40(4), B1101–B1130.
    [Google Scholar]
http://instance.metastore.ingenta.com/content/papers/10.3997/2214-4609.201900870
Loading
/content/papers/10.3997/2214-4609.201900870
Loading

Data & Media loading...

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error