La fédération
Intranet
Séminaire ISTerre
Imaging the Near Surface: Can Full Waveform Inversion and Deep Learning Help?
mardi 12 novembre 2024 - 15h00
Dr. Gabriel Fabien-Ouellet - Assistant Professor, Polytechnique Montréal---
The near surface has a wide societal impact: it is the interface with the built infrastructure and hosts important resources such as groundwater and minerals. Seismic imaging is one of the most powerful techniques to image and monitor the subsurface, supporting applications in geotechnical engineering, hydrogeology, seismic microzonation, and carbon capture and storage monitoring. However, advanced methods like Full Waveform Inversion (FWI), despite their potential, are rarely applied to the near surface due to significant challenges. This presentation explores new perspectives and developments aimed at overcoming these obstacles to make FWI a viable tool for near-surface imaging and monitoring. A key challenge is the high attenuation in unconsolidated or weathered formations, which necessitates computationally intensive multiparameter viscoelastic inversion. One promising approach to reduce the complexity of viscoelastic FWI is the integration of deep learning techniques. We will review recent advancements in applying deep learning within the FWI framework, as well as the key challenges that must be addressed for realistic implementation. To illustrate these advancements, we will showcase the application of deep learning-based seismic inversion for imaging subsea permafrost. Additionally, we will discuss the potential of FWI-guided deep learning for subsurface monitoring, particularly in conjunction with emerging technologies such as Distributed Acoustic Sensing (DAS) and seismic interferometry.
Equipe organisatrice : Ondes et structures
Séminaire uniquement en visio