Inversion of 1D frequency- and time-domain electromagnetic data with convolutional neural networks

Puzyrev, Vladimir; Swidinsky, Andrei

doi:10.1016/j.cageo.2020.104681

Cited by 53 publications

(9 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2, in the generation of the DL-RMD. We also train an additional network using the random resistivity models, similarly to several DL studies (Colombo et al, 2021b;Moghadas, 2020;Moghadas et al, 2020;Noh et al, 2020;Puzyrev and Swidinsky, 2021;Qin et al, 2019;Wu et al, 2021b) as mentioned in Table 1. To have the same level of complexity, the number of layers, depth discretization, and the number of random resistivity models are kept the same as used to train the other two networks for a fair comparison, and the resistivity of each layer is chosen randomly from a log-uniform distribution to take into account the nonlinearity of the forward responses with the resistivity values.…”

Section: Surrogate Forward-modelling Resultsmentioning

confidence: 99%

DL-RMD: a geophysically constrained electromagnetic resistivity model database (RMD) for deep learning (DL) applications

Asif

Foged

Bording

et al. 2023

Earth Syst. Sci. Data

View full text Add to dashboard Cite

Abstract. Deep learning (DL) algorithms have shown incredible potential in many applications. The success of these data-hungry methods is largely associated with the availability of large-scale datasets, as millions of observations are often required to achieve acceptable performance levels. Recently, there has been an increased interest in applying deep learning methods to geophysical applications where electromagnetic methods are used to map the subsurface geology by observing variations in the electrical resistivity of the subsurface materials. To date, there are no standardized datasets for electromagnetic methods, which hinders the progress, evaluation, benchmarking, and evolution of deep learning algorithms due to data inconsistency. Therefore, we present a large-scale electrical resistivity model database (RMD) with a wide variety of geologically plausible and geophysically resolvable subsurface structures for the commonly deployed ground-based and airborne electromagnetic systems. Potentially, the presented database can be used to build surrogate models of well-known processes and to aid in labour-intensive tasks. The geophysically constrained property of this database will not only achieve enhanced performance and improved generalization but, more importantly, incorporate consistency and credibility into deep learning models. We show the effectiveness of the presented database by surrogating the forward-modelling process, and we urge the geophysical community interested in deep learning for electromagnetic methods to utilize the presented database. The dataset is publicly available at https://doi.org/10.5281/zenodo.7260886 (Asif et al., 2022a).

show abstract

Section: Surrogate Forward-modelling Resultsmentioning

confidence: 99%

DL-RMD: a geophysically constrained electromagnetic resistivity model database (RMD) for deep learning (DL) applications

Asif

Foged

Bording

et al. 2023

Earth Syst. Sci. Data

View full text Add to dashboard Cite

show abstract

“…Röth and Tarantola (1994) were amongst the first to solve an inverse problem in this way using a multilayer perceptron neural network and demonstrated an application of inversion of reflection seismic data to obtain single estimates of 1D velocity profiles. Recently, several authors have further explored this approach for directly solving a geophysical inverse problem, making use of convolutional neural networks (Bai et al., 2020; Moghadas, 2020; Puzyrev & Swidinsky, 2021). A drawback of such methods is that, as in the deterministic solution of an inverse problem, they estimate only a single model, typically without accounting for uncertainty in geophysical data, and do not quantify the uncertainty on the predicted model parameters.…”

Section: Introductionmentioning

confidence: 99%

Use of Machine Learning to Estimate Statistics of the Posterior Distribution in Probabilistic Inverse Problems—An Application to Airborne EM Data

Hansen

Finlay

2022

JGR Solid Earth

View full text Add to dashboard Cite

A key challenge in geoscience is that of combining different kinds of geo-information into one geo-model, typically describing the subsurface. This information can be direct information about geological processes, and spatial variability, or it can be indirect information from measurements of properties related to the subsurface, such as geophysical data. Ideally, when such a geo-model has been established, one should be able to quantify information about specific features related to the geo-model, consistent with all information.This integration of geo-information is typically solved using inverse problem theory (Menke, 2012;Tarantola & Valette, 1982a). Fast deterministic methods exist and have been widely used. For such methods, the goal is to obtain one optimal model, such as the simplest possible model, consistent with available information, typically in the form of observed data (Constable et al., 1987;Menke, 2012;Tikhonov, 1963). In practice, in part due to noise on data and model nonlinearities and imperfections, infinitely many models exist that will be consistent with data, and the deterministic approach can in general not account properly for such uncertainty.Probabilistic inversion methods can, in principle, take into account arbitrarily complex information, and integrate the information into one consistent model, as given by the posterior probability distribution. A full analytic expression of the posterior distribution is rarely possible. Instead, sampling methods can be used to generate a

show abstract

“…In the context of geophysics, CNNs have been effective with seismic data, used for applications such as fault detection (e.g., some recent papers include Pochet et al, 2018;Zhang, et al, 2019;Cunha et al, 2020), salt classification (e.g., Waldeland and Solberg, 2017;Shi et al, 2018), and horizon tracking (Yang and Sun, 2020). They have also been used with success in electromagnetics (e.g., Puzyrev and Swidinsky, 2021). In aeromagnetics, Nurindrawati and Sun (2020) use CNNs to estimate the total magnetization direction of anomalies, and Aghaee Rad (2019) applies CNNs to aeromagnetic, gravity, and elevation data to determine geologic lineament locations.…”

Section: Introductionmentioning

confidence: 99%

Convolutional neural networks applied to the interpretation of lineaments in aeromagnetic data

Naprstek

Smith

2022

GEOPHYSICS

View full text Add to dashboard Cite

Parameter estimation in aeromagnetics is an important tool for geological interpretation. Due to aeromagnetic data being highly prevalent around the world, it can often be used to assist in understanding the geology of an area as a whole, or for locating potential areas of further investigation for mineral exploration. Methods that automatically provide information such as the location and depth to the source of anomalies are useful to the interpretation, particularly in areas where a large number of anomalies exist. Unfortunately, many of the current methods rely on high-order derivatives, and are therefore susceptible to noise in the data. Convolution neural networks (CNNs) are a subset of machine learning methods that are well-suited to image processing tasks, and have been shown to be effective at interpreting other geophysical data, such as seismic sections. Following several similar successful approaches, we have developed a CNN methodology for estimating the location and depth of lineament-type anomalies in aeromagnetic maps. To train the CNN model, we utilized a synthetic aeromagnetic data modeler to vary relevant physical parameters, and developed a representative dataset of approximately 1.4 million images. These were then used for training classification CNNs, with each class representing a small range of depth values. We first applied the model to a series of difficult synthetic datasets with varying amounts of noise, comparing the results against the tilt-depth method. We then applied the CNN model to a dataset from north-eastern Ontario, Canada, that contained a dyke with known depth, which was correctly estimated. This method is shown to be robust to noise, and can easily be applied to new datasets using the trained model, which has been made publicly available.

show abstract

Inversion of 1D frequency- and time-domain electromagnetic data with convolutional neural networks

Cited by 53 publications

References 33 publications

DL-RMD: a geophysically constrained electromagnetic resistivity model database (RMD) for deep learning (DL) applications

DL-RMD: a geophysically constrained electromagnetic resistivity model database (RMD) for deep learning (DL) applications

Use of Machine Learning to Estimate Statistics of the Posterior Distribution in Probabilistic Inverse Problems—An Application to Airborne EM Data

Convolutional neural networks applied to the interpretation of lineaments in aeromagnetic data

Contact Info

Product

Resources

About