Seismic interferometry with neural networks
Name
segam2021-3594981.1.pdf
Description
Accepted version
Size
5 MB
Format
Adobe PDF
Checksum (MD5)
8b5fb7a17757e00bdfd454ce036715be
Author(s)
Sun, Hongyu
Demanet, Laurent
Date Issued
September 1, 2021
Journal
First International Meeting for Applied Geoscience & Energy
Publisher
Society of Exploration Geophysicists
Citation
Hongyu Sun, Laurent Demanet; September 1, 2021. "Seismic interferometry with neural networks." Proceedings of the First International Meeting for Applied Geoscience & Energy. First International Meeting for Applied Geoscience & Energy. (pp. pp. 1384-1389). ASME.
Version
Final published version
Abstract
Under the assumptions of diffuse wavefields or energy equipartitioning, theoretical studies showed that the Green’s function can be retrieved from the cross-correlation of ambient noise in seismic interferometry (SI). However, in practice, correlograms are not equal to the empirical Green’s function since the assumptions for correlation-based SI are generally not satisfied in realistic situations. In the framework of supervised learning, we propose to train deep neural networks to overcome two limitations of correlation-based SI: the temporal limitation of passive recordings, and the spatial limitation of the random source distribution. Deep neural networks are trained to implicitly find the relationship between the empirical Green’s function and the correlograms, and then used to extract the correct Green’s function from ambient noise. The input of the network is correlograms (a virtual shot gather) and the desired output is the empirical Green’s function (the active shot gather). Numerical examples show that a deep network aware of the source directionality (through a preliminary beamforming step) can help mitigate some of the challenges associated with inhomogeneous source distributions. In this work, all the numerical examples are based on the retrieval of P-wave reflections at exploration scales, and are conducted on synthetic data. Many precautions are taken to avoid the “learning crime” where the training and testing scenarios are too closely related. We use the CycleGAN architecture in all our numerical experiments.
Terms of Use
Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use.
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1190/segam2021-3594981.1
