Deep learning-based automatic horizon identification from seismic data

Harshit Gupta; Siddhant Pradhan; Rahul Gogia; Seshan Srirangarajan; Jyoti Phirani; Sayan Ranu

doi:10.2118/196087-ms

Deep learning-based automatic horizon identification from seismic data

Harshit Gupta, Siddhant Pradhan, Rahul Gogia, Seshan Srirangarajan, Jyoti Phirani, Sayan Ranu

Civil And Environmental Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution book

6 Citations (Scopus)

Abstract

Horizons in a seismic image are geologically signficant surfaces that can be used for understanding geological structures and stratigraphy models. However, horizon tracking in seismic data is a time consuming and challenging task. Saving geologist's time from this seismic interpretation task is essential given the time constraints for the decision making in the oil & gas industry. We take advantage of the deep convolutional neural networks (CNN) to track the horizons directly from the seismic images. We propose a novel automatic seismic horizon tracking method that can reduce the time needed for interpretation, as well as increase the accuracy for the geologists. We show the performance comparison of the proposed CNN model for different training data set sizes and different methods of balancing the classes.

Original language	English
Title of host publication	Society of Petroleum Engineers - SPE Annual Technical Conference and Exhibition 2019, ATCE 2019
Publisher	Society of Petroleum Engineers (SPE)
ISBN (Electronic)	9781613996638
DOIs	https://doi.org/10.2118/196087-ms
Publication status	Published - 23 Sept 2019
Event	SPE Annual Technical Conference and Exhibition 2019, ATCE 2019 - Calgary, Canada Duration: 30 Sept 2019 → 2 Oct 2019

Publication series

Name	Proceedings - SPE Annual Technical Conference and Exhibition
Volume	2019-September

Conference

Conference	SPE Annual Technical Conference and Exhibition 2019, ATCE 2019
Country/Territory	Canada
City	Calgary
Period	30/09/19 → 2/10/19

Funding

This work was supported by the Schlumberger India Technology Centre Pvt. Ltd. (research grant no. FT/05/255/2018). The authors thank IIT Delhi HPC facility for the computational resources. The F3 and Penobscot data sets used in this work are provided by dGB Earth Sciences. We would also like to acknowledge the use of OpendTect software in our work.

Keywords

machine learning
deep learning
training data
artificial Intelligence
seismic data
reservoir characterization
upstream oil & gas
neural network
seismic processing and interpretation

Access to Document

10.2118/196087-ms

Cite this

Gupta, H., Pradhan, S., Gogia, R., Srirangarajan, S., Phirani, J., & Ranu, S. (2019). Deep learning-based automatic horizon identification from seismic data. In Society of Petroleum Engineers - SPE Annual Technical Conference and Exhibition 2019, ATCE 2019 Article SPE-196087-MS (Proceedings - SPE Annual Technical Conference and Exhibition; Vol. 2019-September). Society of Petroleum Engineers (SPE). https://doi.org/10.2118/196087-ms

@inproceedings{2a61efaa572244279b359beab7b53d83,

title = "Deep learning-based automatic horizon identification from seismic data",

abstract = "Horizons in a seismic image are geologically signficant surfaces that can be used for understanding geological structures and stratigraphy models. However, horizon tracking in seismic data is a time consuming and challenging task. Saving geologist's time from this seismic interpretation task is essential given the time constraints for the decision making in the oil & gas industry. We take advantage of the deep convolutional neural networks (CNN) to track the horizons directly from the seismic images. We propose a novel automatic seismic horizon tracking method that can reduce the time needed for interpretation, as well as increase the accuracy for the geologists. We show the performance comparison of the proposed CNN model for different training data set sizes and different methods of balancing the classes.",

keywords = "machine learning, deep learning, training data, artificial Intelligence, seismic data, reservoir characterization, upstream oil & gas, neural network, seismic processing and interpretation",

author = "Harshit Gupta and Siddhant Pradhan and Rahul Gogia and Seshan Srirangarajan and Jyoti Phirani and Sayan Ranu",

note = "Funding Information: This work was supported by the Schlumberger India Technology Centre Pvt. Ltd. (research grant no. FT/05/255/2018). The authors thank IIT Delhi HPC facility for the computational resources. The F3 and Penobscot data sets used in this work are provided by dGB Earth Sciences. We would also like to acknowledge the use of OpendTect software in our work. Paper Number: SPE-196087-MS Publisher Copyright: {\textcopyright} 2019, Society of Petroleum Engineers; SPE Annual Technical Conference and Exhibition 2019, ATCE 2019 ; Conference date: 30-09-2019 Through 02-10-2019",

year = "2019",

month = sep,

day = "23",

doi = "10.2118/196087-ms",

language = "English",

series = "Proceedings - SPE Annual Technical Conference and Exhibition",

publisher = "Society of Petroleum Engineers (SPE)",

booktitle = "Society of Petroleum Engineers - SPE Annual Technical Conference and Exhibition 2019, ATCE 2019",

address = "United States",

}

Gupta, H, Pradhan, S, Gogia, R, Srirangarajan, S, Phirani, J & Ranu, S 2019, Deep learning-based automatic horizon identification from seismic data. in Society of Petroleum Engineers - SPE Annual Technical Conference and Exhibition 2019, ATCE 2019., SPE-196087-MS, Proceedings - SPE Annual Technical Conference and Exhibition, vol. 2019-September, Society of Petroleum Engineers (SPE), SPE Annual Technical Conference and Exhibition 2019, ATCE 2019, Calgary, Canada, 30/09/19. https://doi.org/10.2118/196087-ms

Deep learning-based automatic horizon identification from seismic data. / Gupta, Harshit; Pradhan, Siddhant; Gogia, Rahul et al.
Society of Petroleum Engineers - SPE Annual Technical Conference and Exhibition 2019, ATCE 2019. Society of Petroleum Engineers (SPE), 2019. SPE-196087-MS (Proceedings - SPE Annual Technical Conference and Exhibition; Vol. 2019-September).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution book

TY - GEN

T1 - Deep learning-based automatic horizon identification from seismic data

AU - Gupta, Harshit

AU - Pradhan, Siddhant

AU - Gogia, Rahul

AU - Srirangarajan, Seshan

AU - Phirani, Jyoti

AU - Ranu, Sayan

N1 - Funding Information: This work was supported by the Schlumberger India Technology Centre Pvt. Ltd. (research grant no. FT/05/255/2018). The authors thank IIT Delhi HPC facility for the computational resources. The F3 and Penobscot data sets used in this work are provided by dGB Earth Sciences. We would also like to acknowledge the use of OpendTect software in our work. Paper Number: SPE-196087-MS Publisher Copyright: © 2019, Society of Petroleum Engineers

PY - 2019/9/23

Y1 - 2019/9/23

N2 - Horizons in a seismic image are geologically signficant surfaces that can be used for understanding geological structures and stratigraphy models. However, horizon tracking in seismic data is a time consuming and challenging task. Saving geologist's time from this seismic interpretation task is essential given the time constraints for the decision making in the oil & gas industry. We take advantage of the deep convolutional neural networks (CNN) to track the horizons directly from the seismic images. We propose a novel automatic seismic horizon tracking method that can reduce the time needed for interpretation, as well as increase the accuracy for the geologists. We show the performance comparison of the proposed CNN model for different training data set sizes and different methods of balancing the classes.

AB - Horizons in a seismic image are geologically signficant surfaces that can be used for understanding geological structures and stratigraphy models. However, horizon tracking in seismic data is a time consuming and challenging task. Saving geologist's time from this seismic interpretation task is essential given the time constraints for the decision making in the oil & gas industry. We take advantage of the deep convolutional neural networks (CNN) to track the horizons directly from the seismic images. We propose a novel automatic seismic horizon tracking method that can reduce the time needed for interpretation, as well as increase the accuracy for the geologists. We show the performance comparison of the proposed CNN model for different training data set sizes and different methods of balancing the classes.

KW - machine learning

KW - deep learning

KW - training data

KW - artificial Intelligence

KW - seismic data

KW - reservoir characterization

KW - upstream oil & gas

KW - neural network

KW - seismic processing and interpretation

UR - http://www.scopus.com/inward/record.url?scp=85079185120&partnerID=8YFLogxK

U2 - 10.2118/196087-ms

DO - 10.2118/196087-ms

M3 - Conference contribution book

AN - SCOPUS:85079185120

T3 - Proceedings - SPE Annual Technical Conference and Exhibition

BT - Society of Petroleum Engineers - SPE Annual Technical Conference and Exhibition 2019, ATCE 2019

PB - Society of Petroleum Engineers (SPE)

T2 - SPE Annual Technical Conference and Exhibition 2019, ATCE 2019

Y2 - 30 September 2019 through 2 October 2019

ER -

Gupta H, Pradhan S, Gogia R, Srirangarajan S, Phirani J, Ranu S. Deep learning-based automatic horizon identification from seismic data. In Society of Petroleum Engineers - SPE Annual Technical Conference and Exhibition 2019, ATCE 2019. Society of Petroleum Engineers (SPE). 2019. SPE-196087-MS. (Proceedings - SPE Annual Technical Conference and Exhibition). doi: 10.2118/196087-ms

Deep learning-based automatic horizon identification from seismic data

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this