Layered low-frequency extrapolation with deep learning in full-waveform inversion

Yao Liu; Baodi Liu; Jianping Huang; Jun Wang; Honglong Chen; Weifeng Liu

doi:10.1117/12.2612130

19 November 2021 Layered low-frequency extrapolation with deep learning in full-waveform inversion

Yao Liu, Baodi Liu, Jianping Huang, Jun Wang, Honglong Chen, Weifeng Liu

Proceedings Volume 12059, Tenth International Symposium on Precision Mechanical Measurements; 1205919 (2021) https://doi.org/10.1117/12.2612130
Event: Tenth International Symposium on Precision Mechanical Measurements, 2021, Qingdao, China

Abstract

The low-frequency information of seismic records can enhance the recognition ability of lithological bodies and make the inversion results clear and reliable. Affected by conventional acquisition technology, low-frequency information is usually missing in the imaging profile. Therefore, extrapolating low-frequency seismic data from band-limited seismic data is an important research topic in full-waveform inversion (FWI). Most of the existing methods directly use machine learning to extrapolate the low-frequency, but the amplitude of the seismic records will be greatly attenuated with the increase of offset and time. The energy gap of seismic records is wide, and the contribution of high-energy seismic records to the network weight is far greater than that of low-energy data. Therefore, it is difficult to extrapolate deep low-energy data. To solve this problem, we propose a method of layered low-frequency extrapolation with deep learning. The seismic records are divided into several layers according to the change of depth and the similarity of energy, and the convolutional neural network is used for training. Experimental results show that this method can accurately extrapolate low-frequency data, and the extrapolation data in the deep layer are close to true data in both the time domain and the frequency domain. In addition, this method occupies lesser computing resources and has the potential for field data application. We verify the effectiveness of the method through two datasets obtained from the Marmousi model and the overthrust model.

Citation Download Citation

Yao Liu, Baodi Liu, Jianping Huang, Jun Wang, Honglong Chen, and Weifeng Liu "Layered low-frequency extrapolation with deep learning in full-waveform inversion", Proc. SPIE 12059, Tenth International Symposium on Precision Mechanical Measurements, 1205919 (19 November 2021); https://doi.org/10.1117/12.2612130

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
8 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Data modeling

Fusion energy

Data fusion

Wavelets

Convolution

Machine learning

Neural networks

Show All Keywords

Keywords/Phrases

Search In:

Publication Years