2020/06/27 by Ahmad Mustafa, Mustafa, Ahmad, Motaz Alfarraj +3 · 1 citation
Earth and Planetary Sciences · Engineering · #Drilling and Well Engineering #FOS: Computer and information sciences #FOS: Electrical engineering #FOS: Physical sciences #Geophysics (physics.geo-ph) #Image and Video Processing (eess.IV) #Machine Learning (stat.ML) #Reservoir Engineering and Simulation Methods #Seismic Imaging and Inversion Techniques #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2006.15472
openalex publication_date 2020/06/27 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
Seismic inversion refers to the process of estimating reservoir rock\nproperties from seismic reflection data. Conventional and machine\nlearning-based inversion workflows usually work in a trace-by-trace fashion on\nseismic data, utilizing little to no information from the spatial structure of\nseismic images. We propose a deep learning-based seismic inversion workflow\nthat models each seismic trace not only temporally but also spatially. This\nutilizes information-relatedness in seismic traces in depth and spatial\ndirections to make efficient rock property estimations. We empirically compare\nour proposed workflow with some other sequence modeling-based neural networks\nthat model seismic data only temporally. Our results on the SEAM dataset\ndemonstrate that, compared to the other architectures used in the study, the\nproposed workflow is able to achieve the best performance, with an average\nr2 coefficient of 79.77 %.\n