Deep learning prior models from seismic images for full-waveform inversion

Lewis, Winston; Vigh, Denes

doi:10.1190/segam2017-17627643.1

Cited by 118 publications

(38 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Araya-Polo et al (2018) develop a formulation for FWI that replaces the iterative inversion scheme for velocity features with a deep NN. Lewis and Vigh (2017) use a deep NN FWI method to better detect salt domes. Such methods appear to be a future step for generative and inversion architectures.…”

Section: Ground-motion Prediction Using Supervised Learningmentioning

confidence: 99%

Machine Learning in Seismology: Turning Data into Insights

Kong

Trugman

Ross

et al. 2018

Seismological Research Letters

397

197

View full text Add to dashboard Cite

This article provides an overview of current applications of machine learning (ML) in seismology. ML techniques are becoming increasingly widespread in seismology, with applications ranging from identifying unseen signals and patterns to extracting features that might improve our physical understanding. The survey of the applications in seismology presented here serves as a catalyst for further use of ML. Five research areas in seismology are surveyed in which ML classification, regression, clustering algorithms show promise: earthquake detection and phase picking, earthquake early warning (EEW), ground-motion prediction, seismic tomography, and earthquake geodesy. We conclude by discussing the need for a hybrid approach combining data-driven ML with traditional physical modeling.

show abstract

Section: Ground-motion Prediction Using Supervised Learningmentioning

confidence: 99%

Machine Learning in Seismology: Turning Data into Insights

Kong

Trugman

Ross

et al. 2018

Seismological Research Letters

397

197

View full text Add to dashboard Cite

show abstract

“…Recent ANN applications to subsurface imaging claim to overcome these deficiencies, using seismic data as input to identify important structures. In particular, [39] uses earthquake data to accurately predict 1-D velocity models, and applications in [5,34,27,53] employ data collected in seismic surveys for structural model building with interest in hydrocarbon exploration. In addition, the study in [37] applies a ANN to infer the prior distribution of acoustic properties of a geological model, that is later improved by full waveform inversion.…”

Section: Machine Learning Applications To Earthquake Datamentioning

confidence: 99%

Artificial Neural Networks as Emerging Tools for Earthquake Detection

Rojas

Otero

Alvarado

et al. 2019

CyS

View full text Add to dashboard Cite

As seismic networks continue to spread and monitoring sensors become more efficient, the abundance of data highly surpasses the processing capabilities of earthquake interpretation analysts. Earthquake catalogs are fundamental for fault system studies, event modellings, seismic hazard assessment, forecasting, and ultimately, for mitigating the seismic risk. These have fueled the research for the automation of interpretation tasks such as event detection, event identification, hypocenter location, and source mechanism analysis. Over the last forty years, traditional algorithms based on quantitative analyses of seismic traces in the time or frequency domain, have been developed to assist interpretation. Alternatively, recent advances are related to the application of Artificial Neural Networks (ANNs), a subset of machine learning techniques that is pushing the state-of-the-art forward in many areas. Appropriated trained ANN can mimic the interpretation abilities of best human analysts, avoiding the individual weaknesses of most traditional algorithms, and spending modest computational resources at the operational stage. In this paper, we will survey the latest ANN applications to the automatic interpretation of seismic data, with a special focus on earthquake detection, and the estimation of onset times. For a comparative framework, we give an insight into the labor of human interpreters, who may face uncertainties in the case of small magnitude earthquakes.

show abstract

“…The testing performance showed that salt detections is much faster and efficient by this method than traditional migration and interpretation. Lewis and Vigh (2017) investigated a combination of DL and FWI to improve the performance for salt inversion. In that study, the network was trained to generate useful prior models for FWI by learning features relevant to earth model building from a seismic image.…”

Section: Introductionmentioning

confidence: 99%

Deep-learning inversion: A next-generation seismic velocity model building method

2019

View full text Add to dashboard Cite

Seismic velocity is one of the most important parameters used in seismic exploration.Accurate velocity models are key prerequisites for reverse-time migration and other high-resolution seismic imaging techniques. Such velocity information has traditionally been derived by tomography or full-waveform inversion (FWI), which are time consuming and computationally expensive, and they rely heavily on human interaction and quality control. We investigate a novel method based on the supervised deep fully convolutional neural network (FCN) for velocity-model building (VMB) directly from raw seismograms. Unlike the conventional inversion method based on physical models, the supervised deep-learning methods are based on big-data training rather than prior-knowledge assumptions. During the training stage, the network establishes a nonlinear projection from the multi-shot seismic data to the corresponding velocity models.During the prediction stage, the trained network can be used to estimate the velocity models from the new input seismic data. One key characteristic of the deep-learning method is that it can automatically extract multi-layer useful features without the need for human-curated activities and initial velocity setup. The data-driven method usually requires more time during the training stage, and actual predictions take less time, with only seconds needed. Therefore, the computational time of geophysical inversions, including real-time inversions, can be dramatically reduced once a good generalized network is built. By using numerical experiments on synthetic models, the promising performances of our proposed method are shown in comparison with conventional FWI even when the input data are in more realistic scenarios. Discussions on the deep-learning methods, training dataset, lack of low frequencies, and advantages and disadvantages of the new method are also provided.

show abstract

Deep learning prior models from seismic images for full-waveform inversion

Cited by 118 publications

References 15 publications

Machine Learning in Seismology: Turning Data into Insights

Machine Learning in Seismology: Turning Data into Insights

Artificial Neural Networks as Emerging Tools for Earthquake Detection

Deep-learning inversion: A next-generation seismic velocity model building method

Contact Info

Product

Resources

About