Semi‐supervised deep autoencoder for seismic facies classification

Liu, Xingye; Li, Bin; Li, Jingye; Chen, Xiaohong; Li, Qingchun; Chen, Yangkang

doi:10.1111/1365-2478.13106

Cited by 19 publications

(10 citation statements)

References 63 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Deep learning is a data‐driven algorithm that can learn implicit nonlinear relations from label data and use it to solve nonlinear problems. Deep learning has been widely used in geophysical problems such as seismic facies analysis (Liu et al., 2021; Nishitsuji & Exley, 2019), first‐break picking (Wang et al., 2019; Yuan et al., 2018; Zwartjes & Yoo, 2022), fault identification (Huang et al., 2017; Wu et al., 2019; Zhou et al., 2021) and model building (Araya‐Polo et al., 2017; Fabien‐Ouellet & Sarkar, 2019; Ovcharenko et al., 2022). Recently, the application of deep neural networks in reservoir characterization has also been investigated (Chen & Saygin, 2021; Dhara et al., 2023; Di & Abubakar, 2021; Sun et al., 2021; Wu et al., 2021).…”

Section: Introductionmentioning

confidence: 99%

Building initial model for seismic inversion based on semi‐supervised learning

Sun,

Zong

2024

Geophysical Prospecting

View full text Add to dashboard Cite

Seismic inversion is an important tool for reservoir characterization. The inversion results are significantly impacted by a reliable initial model. Conventional well interpolation methods are not able to meet the needs of seismic inversion for lateral heterogeneous reservoirs. Inspired by the sequence modelling network and seismic inversion in the Laplace–Fourier domain, we propose an initial model‐building method using semi‐supervised learning strategy. The proposed method considers spatial information to ensure the horizontal continuity of the initial model. Based on the fact that the low‐frequency components of seismic signals in the Laplace–Fourier domain are easier to obtain, we use the forward model in the Laplace–Fourier domain to replace the time‐domain forward model. The proposed workflow was validated using the Marmousi II model. Although the training was carried out on a small number of low‐frequency impedance traces, the proposed workflow was able to build low‐frequency model for the entire Marmousi II model with a correlation of 98%. Field data examples demonstrate the feasibility and effectiveness of the proposed method. For lateral heterogeneous reservoirs, the proposed method performs better than the well interpolation method. By utilizing the model obtained by the proposed method as the initial low‐frequency model of the conventional inversion method, it is possible to estimate better inversion results. The results of different combinations of training sets demonstrate the stability of the proposed method. This method may still be a viable choice if there is lateral heterogeneity underground but not much well‐logging label data.

show abstract

Section: Introductionmentioning

confidence: 99%

Building initial model for seismic inversion based on semi‐supervised learning

Sun,

Zong

2024

Geophysical Prospecting

View full text Add to dashboard Cite

show abstract

“…Many existing studies have focused on the application of individual methods without fully exploring the synergistic potential of combining various computational algorithms (X. Liu et al, 2021). This research aims to address this gap by providing an integrated, multifaceted approach that leverages the strengths of various computational techniques in unison.…”

Section: Introductionmentioning

confidence: 99%

Advanced Seismic Data Analysis: Comparative study of Machine Learning and Deep Learning for Data Prediction and Understanding

Airlangga

2024

Brilliance

View full text Add to dashboard Cite

This study delves into the application of machine learning (ML) and deep learning (DL) techniques for the analysis of seismic data, aiming to identify and categorize patterns and anomalies within seismic events. Using a robust dataset, we applied three distinct clustering approaches: K-Means, DBSCAN, and an Autoencoder-based method, each offering unique perspectives on the data. K-Means clustering provided a fundamental partitioning of the data into five predefined clusters, facilitating the identification of broad seismic patterns. DBSCAN, a density-based clustering algorithm, offered insights into the spatial distribution and density of seismic events, adeptly pinpointing anomalies and outliers that signify unusual seismic activity. The Autoencoder, leveraging deep learning, excelled in capturing complex and non-linear relationships within the data, revealing subtle patterns not immediately apparent through traditional methods. The effectiveness of these clustering techniques was quantitatively evaluated using the Silhouette Score and the Davies-Bouldin Score, alongside visual assessments through PCA and t-SNE for dimensionality reduction. The results indicated that while K-Means provided clear partitioning, DBSCAN excelled in outlier detection, and the Autoencoder offered a balanced approach with its nuanced analysis capabilities. Our comprehensive analysis underscores the significance of employing a multi-methodological approach in seismic data analysis, as each method contributes uniquely to the understanding of seismic events. The insights gained from this study are valuable for enhancing predictive models and improving disaster risk management strategies in seismology. Future research directions include the integration of additional seismic features, validation against larger datasets, and the development of hybrid models to further refine the predictive accuracy of seismic event analysis.

show abstract

“…Over a period of several decades, deep learning (DL) has been developing rapidly (LeCun et al ., 2015), which has brought increased attention from geophysicists to apply the DL strategy to problems related to seismic exploration (Yu and Ma, 2021), such as velocity model building (Yang and Ma, 2019; Li et al ., 2020; Sun et al ., 2021; bin Waheed et al ., 2021), fault detection (Wu et al ., 2019b), noise attenuation (Gao et al ., 2021; Sang et al ., 2021; Yang et al ., 2021a, 2021b), facies classification (Liu et al ., 2021), trace interpolation (Wang et al ., 2020a), lithofacies prediction (Zhao et al ., 2021), permeability and porosity prediction (Yang et al ., 2022), velocity picking (Wang et al ., 2021b) and full‐waveform inversion (Zhang and Alkhalifah, 2019; Huang and Zhu, 2020). Considerable research has been carried out previously for the development of DL‐based methods for seismic inversion; one of the simplest and most commonly adopted methods is to train a supervised network using a mass of seismic data and the corresponding impedance model and then input all the seismic data into the trained network architecture to predict the full impedance.…”

Section: Introductionmentioning

confidence: 99%

Seismic impedance inversion using a multi‐input neural network with a two‐step training strategy

Meng

Wang

Niu

et al. 2022

Geophysical Prospecting

View full text Add to dashboard Cite

Deep learning has shown excellent performance in simulating complex nonlinear mappings from the seismic data to elastic parameters. However, seismic acoustic impedance estimated from a direct mapping from seismic waveform data to P‐wave impedance (single‐input network) is hampered by the limited frequency bands. In this paper, we propose to incorporate the low‐frequency impedance model to constrain the inversion (multi‐input network). We add a feature fusion layer to force the lateral smoothness. Besides, usually, a given seismic survey is likely to contain only a few well logs, which is insufficient for conventional deep‐learning ‐based methods to learn the complex mapping from seismic data to elastic parameters. The problem is compounded by the fact that a network trained with synthetic data (compensated for the lack of logs) cannot be directly used for field data. Therefore, we propose to use transfer learning to mitigate this issue. The multi‐input neural network is trained using synthetics and real data in two stages. We carry out experiments to demonstrate that the two‐step training multi‐input network approach has high accuracy in the time direction, excellent continuity in the lateral direction and favourable robustness. Synthetic and field data examples demonstrate that the proposed network can accurately predict impedance even with limited logging data, which provides a reference for oil and gas exploration in the actual production process.

show abstract

Semi‐supervised deep autoencoder for seismic facies classification

Cited by 19 publications

References 63 publications

Building initial model for seismic inversion based on semi‐supervised learning

Building initial model for seismic inversion based on semi‐supervised learning

Advanced Seismic Data Analysis: Comparative study of Machine Learning and Deep Learning for Data Prediction and Understanding

Seismic impedance inversion using a multi‐input neural network with a two‐step training strategy

Contact Info

Product

Resources

About