Article (Scientific journals)
Geophysical Inversion Using a Variational Autoencoder to Model an Assembled Spatial Prior Uncertainty
Lopez-Alvis, J.; Nguyen, Frédéric; Looms, M.C. et al.
2022In Journal of Geophysical Research. Solid Earth, 127 (3)
Peer Reviewed verified by ORBi
 

Files


Full Text
Lopez_Alvis_etal_2022_JGRSE.pdf
Author postprint (3.58 MB)
Download

All documents in ORBi are protected by a user license.

Send to



Details



Keywords :
deep learning; geophysical inversion; ground-penetrating radar; prior information; traveltime tomography; variational autoencoder; Geophysics; Geochemistry and Petrology; Space and Planetary Science; Earth and Planetary Sciences (miscellaneous)
Abstract :
[en] Prior information regarding subsurface spatial patterns may be used in geophysical inversion to obtain realistic subsurface models. Field experiments require prior information with sufficiently diverse patterns to accurately estimate the spatial distribution of geophysical properties in the sensed subsurface domain. A variational autoencoder (VAE) provides a way to assemble all patterns deemed possible in a single prior distribution. Such patterns may include those defined by different base training images and also their perturbed versions, for example, those resulting from geologically consistent operations such as erosion/dilation, local deformation, and intrafacies variability. Once the VAE is trained, inversion may be done in the latent space which ensures that inverted models have the patterns defined by the assembled prior. Gradient-based inversion with both a synthetic and a field case of cross-borehole GPR traveltime data shows that using the VAE assembled prior performs as good as using the VAE trained on the pattern with the best fit, but it has the advantage of lower computation cost and more realistic prior uncertainty. Moreover, the synthetic case shows an adequate estimation of most small-scale structures. The absolute values of wave velocity are computed by assuming a linear mixing model which involves two additional parameters that effectively shift and scale velocity values and are included in the inversion.
Disciplines :
Geological, petroleum & mining engineering
Author, co-author :
Lopez-Alvis, J. ;  Urban and Environmental Engineering, Applied Geophysics, University of Liege, Liege, Belgium
Nguyen, Frédéric  ;  Université de Liège - ULiège > Département ArGEnCo > Géophysique appliquée
Looms, M.C. ;  Department of Geosciences and Natural Resource Management, University of Copenhagen, Copenhagen, Denmark
Hermans, T. ;  Department of Geology, Ghent University, Gent, Belgium
Language :
English
Title :
Geophysical Inversion Using a Variational Autoencoder to Model an Assembled Spatial Prior Uncertainty
Publication date :
March 2022
Journal title :
Journal of Geophysical Research. Solid Earth
ISSN :
2169-9313
eISSN :
2169-9356
Publisher :
John Wiley and Sons Inc
Volume :
127
Issue :
3
Peer reviewed :
Peer Reviewed verified by ORBi
European Projects :
H2020 - 722028 - ENIGMA - European training Network for In situ imaGing of dynaMic processes in heterogeneous subsurfAce environments
Funders :
EU - European Union [BE]
Funding text :
This work has received funding from the European Unions Horizon 2020 research and innovation program under the Marie Sklodowska‐Curie grant agreement number 722 028 (ENIGMA ITN). We thank two anonymous reviewers and the associate editor for their valuable comments that greatly improved the manuscript.
Available on ORBi :
since 22 November 2023

Statistics


Number of views
32 (3 by ULiège)
Number of downloads
11 (1 by ULiège)

Scopus citations®
 
15
Scopus citations®
without self-citations
13
OpenCitations
 
9

Bibliography


Similar publications



Contact ORBi