Deep Learning Ensemble for Seismic First-Break Event Picking

Zhao, T.; Bilsby, Phillip; Manikani, Sunil; Busanello, G.; Benzaoui, M.; Abubakar, Aria

doi:10.3997/2214-4609.202210425

Cited by 2 publications

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…By applying the MC PM process to synthetic data, we have shown that, in most cases, t e provides the most significant contribution to the overall uncertainty in z N , of the three errors explored. Since t e is the most significant source of uncertainty in an output model, field effort should be directed towards maximising SNR -i.e., through the use of more energetic seismic sources and short maximum source-geophone offsets, minimising wind noise on geophones, stacking more sources at any shotpoint, and/or through applying advanced first-break picking techniques (e.g., Zhao et al, 2022). In addition to minimising t e , efforts should be made to obtain accurate estimates of the magnitude of this error to provide realistic estimates of output uncertainties (e.g., Abakumov et al, 2020).…”

Section: Discussionmentioning

confidence: 99%

The sensitivity of seismic refraction velocity models to survey geometry errors, assessed using Monte Carlo analysis

Watts

Booth

Clark

et al. 2023

Journal of Applied Geophysics

View full text Add to dashboard Cite

Section: Discussionmentioning

confidence: 99%

The sensitivity of seismic refraction velocity models to survey geometry errors, assessed using Monte Carlo analysis

Watts

Booth

Clark

et al. 2023

Journal of Applied Geophysics

View full text Add to dashboard Cite

Improving the accuracy of predicted first-arrival times using deep learning

Pu,

Zhang,

Wei

et al. 2024

GEOPHYSICS

View full text Add to dashboard Cite

Recently, the task of first-arrival time picking for seismic shot gathers is treated as an image segmentation problem and deep learning (DL) algorithms have been successfully used to predict the first-arrival times. Currently, researchers mainly focus on leveraging cutting-edge DL algorithms to improve the performance of DL in first-arrival picking. There are few publications addressing quality control of results predicted by DL. We propose a three-step workflow to improve the accuracy of first-arrival time detection computed using DL algorithms. First, we obtain three predicted results (generation I) by applying the Historically nested U-Net (HU-net) to seismic shot gathers, the envelope of seismic shot gathers, and the cosine of the instantaneous phase of seismic shot gathers. Subsequently, we obtain generation II picking by statistically analyzing the predicted generation I picking. Finally, we treat the first-arrival picking task as a constrained path search problem and the generation II picking function as the constraints. The proposed workflow is applied to real seismic surveys to demonstrate its effectiveness.

show abstract

Deep Learning Ensemble for Seismic First-Break Event Picking

Cited by 2 publications

References 0 publications

The sensitivity of seismic refraction velocity models to survey geometry errors, assessed using Monte Carlo analysis

The sensitivity of seismic refraction velocity models to survey geometry errors, assessed using Monte Carlo analysis

Improving the accuracy of predicted first-arrival times using deep learning

Contact Info

Product

Resources

About