eduzhai > Applied Sciences > Engineering >

Spatiotemporal Modeling of Seismic Images for Acoustic Impedance Estimation

  • Save

... pages left unread,continue reading

Document pages: 6 pages

Abstract: Seismic inversion refers to the process of estimating reservoir rockproperties from seismic reflection data. Conventional and machinelearning-based inversion workflows usually work in a trace-by-trace fashion onseismic data, utilizing little to no information from the spatial structure ofseismic images. We propose a deep learning-based seismic inversion workflowthat models each seismic trace not only temporally but also spatially. Thisutilizes information-relatedness in seismic traces in depth and spatialdirections to make efficient rock property estimations. We empirically compareour proposed workflow with some other sequence modeling-based neural networksthat model seismic data only temporally. Our results on the SEAM datasetdemonstrate that, compared to the other architectures used in the study, theproposed workflow is able to achieve the best performance, with an average$r^{2}$ coefficient of 79.77 .

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×