Deep learning and its application in geochemical mapping

Deep learning for seismic lithology prediction

In this letter, we use a fully convolutional residual network (FCRN) for seismic impedance inversion. After training with appropriate data, the FCRN can effectively predict impedance with high accuracy, and have good robustness against noise and phase difference. However, it cannot give acceptable results in training and predicting models with different geological features. Transfer learning is later introduced to ease this problem. Marmousi2 and Overthrust models are used to verify the effectiveness of the proposed method. Tests show that after fine-tuned by five traces of Overthrust model, the FCRN trained on the Marmousi2 model can give a comparable result similarly predicted by the FCRN trained purely on the Overthrust model.

Seismic Impedance Inversion Using Fully Convolutional Residual Network and Transfer Learning

Velocity analysis can be a time-consuming task when performed manually. Methods have been proposed to automate the process of velocity analysis, which, however, typically requires significa...

Automatic velocity analysis using convolutional neural network and transfer learning

Seismic inversion is a process of predicting high-resolution stratigraphic parameters from low-resolution seismic data. Traditional inversion methods tend to impose human prior knowledge, such as sparsity, to the modeling of the seismic inversion process. Nowadays, with the development of deep learning, the idea of modeling by learning from data has gained great attention in varieties of research fields. As a data-driven method, an artificial neural network (ANN) has already been explored by many researchers in the field of seismic inversion. Compared to ANN, a convolutional neural network (CNN) has a stronger learning ability attributing to its sophisticated structures. However, the development of CNN is limited by the amount of labeled data in many industrial fields including the field of seismic inversion. In order to mitigate the dependence of CNN on the amount of labeled data, we propose a closed-loop CNN structure in this article. The proposed closed-loop CNN can model the seismic forward and inversion process simultaneously from the training data set. Compared to traditional CNN, which is in an open-loop form, closed-loop CNN can not only learn from labeled data but also extract information contained in unlabeled data. The experimental results show that the closed-loop CNN has a better performance than both traditional methods and other deep learning-based methods on the synthetic data set and also can be efficiently applied on the real seismic data set.

Well-Logging Constrained Seismic Inversion Based on Closed-Loop Convolutional Neural Network

Regardless of successful applications of the convolutional neural networks (CNNs) in different fields, its application to seismic waveform classification and first-break (FB) picking has not been explored yet. This letter investigates the application of CNNs for classifying time-space waveforms from seismic shot gathers and picking FBs of both direct wave and refracted wave. We use representative subimage samples with two types of labeled waveform classification to supervise CNNs training. The goal is to obtain the optimal weights and biases in CNNs, which are solved by minimizing the error between predicted and target label classification. The trained CNNs can be utilized to automatically extract a set of time-space attributes or features from any subimage in shot gathers. These attributes are subsequently inputted to the trained fully connected layer of CNNs to output two values between 0 and 1. Based on the two-element outputs, a discriminant score function is defined to provide a single indication for classifying input waveforms. The FB is then located from the calculated score maps by sequentially using a threshold, the first local minimum rule of every trace and a median filter. Finally, we adopt synthetic and real shot data examples to demonstrate the effectiveness of CNNs-based waveform classification and FB picking. The results illustrate that CNN is an efficient automatic data-driven classifier and picker.

/pdf/seismic-waveform-classification-and-first-break-picking-1776zxtuxs.pdf

Seismic Waveform Classification and First-Break Picking Using Convolution Neural Networks

Time-frequency (TF) analysis is a useful tool for seismic data processing and interpretation. We introduce sparse Bayesian learning (SBL) to TF analysis and propose a new SBL-based high-resolution TF method. The method decomposes the seismic trace into a series of Ricker wavelets using SBL-based sparse representations and subsequently implements Wigner–Ville distribution (WVD) on the decomposed wavelets to produce TF spectra. By iteratively solving a Bayesian maximum posterior and a type-II maximum likelihood, SBL-based decomposition can sequentially obtain an optimal number of Ricker wavelets with different peak frequencies or phases from a preset wavelet dictionary, and can simultaneously invert for the associated sparse TF pseudoreflectivity with the prediction uncertainty. The WVD of SBL-based decomposed wavelets can assemble TF distribution of the reconstructed signals to approximately characterize WVD of the original data. Therefore, the linear stack of WVD of all decomposed independent wavelets is immune from both the notorious cross-term interferences of the traditional WVD and random noise. Synthetic data example involving thin beds and laboratorial physical modeling data example involving several known multicave combinations are used to demonstrate the effectiveness of the proposed SBL-based TF analysis method and illustrate its advantages over WVD and the orthogonal matching pursuit-based TF analysis method. The 3-D real seismic data example is adopted to test its application potential for interpreting deep channels and the karst slope fracture zone. The results show that the proposed SBL-based TF method is a potentially effective, stable and high-resolution seismic TF analysis tool even in the presence of thin beds.

Sparse Bayesian Learning-Based Seismic High-Resolution Time-Frequency Analysis

AVAZ inversion for fracture weakness based on three-term Rüger equation


 Robust automatic event detection and location is central to real-time earthquake monitoring. With the increase of computing power and data availability, automated workflows that utilize machine learning (ML) techniques have become increasingly popular; however, ML-based classical workflows still face challenges when applied to the analysis of microseismic data. These seismic sequences are often characterized by short interevent times and/or low signal-to-noise ratio (SNR). Full waveform methods that do not rely on phase picking and association are suitable for processing such datasets, but are computationally costly and lack clear event identification criteria, which is not ideal for real-time processing. To leverage the advantages of both the methods, we propose a new workflow—MAchine Learning aided earthquake MIgration location (MALMI), which integrates ML and waveform migration to perform automated event detection and location. The new workflow uses a pretrained ML model to generate continuous phase probabilities that are then backprojected and stacked to locate seismic sources using migration. We applied the workflow to one month of continuous data collected in the Hengill geothermal area of Iceland to monitor induced earthquakes around two geothermal production sites. With a ML model (EQ-Transformer) pretrained using a global distribution of earthquakes, the proposed workflow automatically detects and locates 250 additional seismic events (accounting for 36% events in the obtained catalog) compared to a reference catalog generated using the SeisComP software. Most of the new events are microseismic events with a magnitude less than 0. Visual inspection of the waveforms of the newly detected events indicates that they are real seismic events of low SNR and are only reliably recorded by very few stations in the array. Further comparison with the conventional migration method based on short-term average over long-term average confirms that MALMI can produce much clearer stacked images with higher resolution and reliability, especially for events with low SNR. The workflow is freely available on GitHub, providing an automated tool for simultaneous event detection and location from continuous seismic data.

https://zenodo.org/record/6822864/files/Shi_etal_2022_SRL.pdf

MALMI: An Automated Earthquake Detection and Location Workflow Based on Machine Learning and Waveform Migration

Detection and identification of subsurface anomalous structures are key objectives in seismic exploration. The coherence technique has been successfully used to identify geologic abnormalities and discontinuities, such as faults and unconformities. Based on the classic third eigenvalue-based coherence (C3) algorithm, we make several improvements and develop a new method to construct covariance matrix using the original and Hilbert transformed seismic traces. This new covariance matrix more readily converges to the main effective signal energy on the largest eigenvalue by decreasing all other eigenvalues. Compared with the conventional coherence algorithms, our algorithm has higher resolution and better noise immunity ability. Next, we incorporate this new eigenvalue-based algorithm with time-lag dip scanning to relieve the dip effect and highlight the discontinuities. Application on 2D synthetic data demonstrates that our coherence algorithm favorably alleviates the low-valued artifacts caused by ...

Peidong Shi

Papers

Seismic Waveform Classification and First-Break Picking Using Convolution Neural Networks

Sparse Bayesian Learning-Based Seismic High-Resolution Time-Frequency Analysis

AVAZ inversion for fracture weakness based on three-term Rüger equation

MALMI: An Automated Earthquake Detection and Location Workflow Based on Machine Learning and Waveform Migration

Improved eigenvalue-based coherence algorithm with dip scanning