Processing and inversion of landstreamer data using deep learning for near-surface seismic imaging

Seismic techniques are used to image the structure of the subsurface and expand our knowledge about the geological condition of the underground. The information acquired from these techniques is employed in domains such as the petroleum industry and clean energy development. Seismic studies also play a vital role in geotechnical projects and civil engineering. The acquired seismic data must be processed to be interpretable. Conventional methods for processing seismic data are time-consuming and labor-intensive. This problem is more noticeable when the data acquisition is performed using a landstreamer for near-surface studies. The near-surface studies are used to image the shallow subsurface for geotechnical purposes. Conventional methods for near-surface imaging are limited to one specific type of seismic data (surface waves, refractions, or reflections). To address these problems, Polytechnique of Montreal and Geostack will collaborate to adopt artificial intelligence (AI) to automate the processing of the seismic data acquired by a landstreamer. In this study, deep learning (DL) methods will be employed to also perform a joint analysis of surface waves, refractions, and reflections. This technique will reduce the processing time significantly while it can lead to a higher resolution and more accurate estimate of the subsurface.

Faculty Supervisor:

Gabriel Fabien-Ouellet

Student:

Partner:

Geostack

Discipline:

Earth science

Sector:

Professional, scientific and technical services

University:

Polytechnique Montréal

Program: