A back propagation neural network for identifying first-break times on cross borehole ground penetrating radar traces

Dale F. Rucker, Paul A Ferre, Mary M Poulton

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Manually picking the first arrival of energy in a series of cross borehole GPR ray traces can be time consuming and subjective, especially when large data sets need to be processed. One possible remedy is the application of a back propagating neural network. Neural network applications have been used previously in seismic studies to pick the arrival of the P and S waves (Dai and MacBeth, 1997; McCormack et al. 1993; Murat et al. 1992). One particular method, which applied a moving window over the trace, is used here with slight modification. Noisy time-amplitude records were first normalized to range from -1 and 1. These data were then filtered such that values between -1 and a negative threshold were set to -1, values between 1 and a positive threshold were set to 1 and all other values were set to zero. The filtered wave was fed through a neural network that searched for a pattern related to a first arrival. Several filtering parameters were tested, including the size of the moving window, the values of the positive and negative thresholds, and neural network parameters pertaining to training and testing. With minimal training, the neural network performed very well compared to hand picking of arrival times on large data sets.

Original languageEnglish (US)
Title of host publicationProceedings of SPIE - The International Society for Optical Engineering
Pages630-634
Number of pages5
Volume4758
DOIs
StatePublished - 2002
Event9th International Conference on Ground Penetrating Radar - Santa Barbara, CA, United States
Duration: Apr 29 2002May 2 2002

Other

Other9th International Conference on Ground Penetrating Radar
CountryUnited States
CitySanta Barbara, CA
Period4/29/025/2/02

Fingerprint

ground penetrating radar
boreholes
Boreholes
Backpropagation
Radar
arrivals
Neural networks
thresholds
education
Ground penetrating radar systems
P waves
S waves
rays
Testing

Keywords

  • Cross-borehole
  • First break pick
  • Neural network

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Condensed Matter Physics

Cite this

Rucker, D. F., Ferre, P. A., & Poulton, M. M. (2002). A back propagation neural network for identifying first-break times on cross borehole ground penetrating radar traces. In Proceedings of SPIE - The International Society for Optical Engineering (Vol. 4758, pp. 630-634) https://doi.org/10.1117/12.462226

A back propagation neural network for identifying first-break times on cross borehole ground penetrating radar traces. / Rucker, Dale F.; Ferre, Paul A; Poulton, Mary M.

Proceedings of SPIE - The International Society for Optical Engineering. Vol. 4758 2002. p. 630-634.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Rucker, DF, Ferre, PA & Poulton, MM 2002, A back propagation neural network for identifying first-break times on cross borehole ground penetrating radar traces. in Proceedings of SPIE - The International Society for Optical Engineering. vol. 4758, pp. 630-634, 9th International Conference on Ground Penetrating Radar, Santa Barbara, CA, United States, 4/29/02. https://doi.org/10.1117/12.462226
Rucker DF, Ferre PA, Poulton MM. A back propagation neural network for identifying first-break times on cross borehole ground penetrating radar traces. In Proceedings of SPIE - The International Society for Optical Engineering. Vol. 4758. 2002. p. 630-634 https://doi.org/10.1117/12.462226
Rucker, Dale F. ; Ferre, Paul A ; Poulton, Mary M. / A back propagation neural network for identifying first-break times on cross borehole ground penetrating radar traces. Proceedings of SPIE - The International Society for Optical Engineering. Vol. 4758 2002. pp. 630-634
@inproceedings{87210ce1701c45b2b2cabd4a951d35e7,
title = "A back propagation neural network for identifying first-break times on cross borehole ground penetrating radar traces",
abstract = "Manually picking the first arrival of energy in a series of cross borehole GPR ray traces can be time consuming and subjective, especially when large data sets need to be processed. One possible remedy is the application of a back propagating neural network. Neural network applications have been used previously in seismic studies to pick the arrival of the P and S waves (Dai and MacBeth, 1997; McCormack et al. 1993; Murat et al. 1992). One particular method, which applied a moving window over the trace, is used here with slight modification. Noisy time-amplitude records were first normalized to range from -1 and 1. These data were then filtered such that values between -1 and a negative threshold were set to -1, values between 1 and a positive threshold were set to 1 and all other values were set to zero. The filtered wave was fed through a neural network that searched for a pattern related to a first arrival. Several filtering parameters were tested, including the size of the moving window, the values of the positive and negative thresholds, and neural network parameters pertaining to training and testing. With minimal training, the neural network performed very well compared to hand picking of arrival times on large data sets.",
keywords = "Cross-borehole, First break pick, Neural network",
author = "Rucker, {Dale F.} and Ferre, {Paul A} and Poulton, {Mary M}",
year = "2002",
doi = "10.1117/12.462226",
language = "English (US)",
volume = "4758",
pages = "630--634",
booktitle = "Proceedings of SPIE - The International Society for Optical Engineering",

}

TY - GEN

T1 - A back propagation neural network for identifying first-break times on cross borehole ground penetrating radar traces

AU - Rucker, Dale F.

AU - Ferre, Paul A

AU - Poulton, Mary M

PY - 2002

Y1 - 2002

N2 - Manually picking the first arrival of energy in a series of cross borehole GPR ray traces can be time consuming and subjective, especially when large data sets need to be processed. One possible remedy is the application of a back propagating neural network. Neural network applications have been used previously in seismic studies to pick the arrival of the P and S waves (Dai and MacBeth, 1997; McCormack et al. 1993; Murat et al. 1992). One particular method, which applied a moving window over the trace, is used here with slight modification. Noisy time-amplitude records were first normalized to range from -1 and 1. These data were then filtered such that values between -1 and a negative threshold were set to -1, values between 1 and a positive threshold were set to 1 and all other values were set to zero. The filtered wave was fed through a neural network that searched for a pattern related to a first arrival. Several filtering parameters were tested, including the size of the moving window, the values of the positive and negative thresholds, and neural network parameters pertaining to training and testing. With minimal training, the neural network performed very well compared to hand picking of arrival times on large data sets.

AB - Manually picking the first arrival of energy in a series of cross borehole GPR ray traces can be time consuming and subjective, especially when large data sets need to be processed. One possible remedy is the application of a back propagating neural network. Neural network applications have been used previously in seismic studies to pick the arrival of the P and S waves (Dai and MacBeth, 1997; McCormack et al. 1993; Murat et al. 1992). One particular method, which applied a moving window over the trace, is used here with slight modification. Noisy time-amplitude records were first normalized to range from -1 and 1. These data were then filtered such that values between -1 and a negative threshold were set to -1, values between 1 and a positive threshold were set to 1 and all other values were set to zero. The filtered wave was fed through a neural network that searched for a pattern related to a first arrival. Several filtering parameters were tested, including the size of the moving window, the values of the positive and negative thresholds, and neural network parameters pertaining to training and testing. With minimal training, the neural network performed very well compared to hand picking of arrival times on large data sets.

KW - Cross-borehole

KW - First break pick

KW - Neural network

UR - http://www.scopus.com/inward/record.url?scp=0036059377&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036059377&partnerID=8YFLogxK

U2 - 10.1117/12.462226

DO - 10.1117/12.462226

M3 - Conference contribution

AN - SCOPUS:0036059377

VL - 4758

SP - 630

EP - 634

BT - Proceedings of SPIE - The International Society for Optical Engineering

ER -