Seismic Fault Detection Using Convolutional and Transformer-Based Models
DOI:
https://doi.org/10.63282/3117-5481/AIJCST-V4I2P103Keywords:
Seismic Fault Detection, Convolutional Neural Networks (Cnns), Transformer Models, Deep Learning, 3D Seismic Interpretation, Hybrid ArchitecturesAbstract
Seismic fault detection is a critical component of subsurface characterization, reservoir evaluation, and geohazard analysis. Traditional approaches including manual interpretation, edge detection filters, and seismic attributes have contributed significantly to structural mapping but remain limited by subjectivity, sensitivity to noise, and challenges in handling complex 3D seismic volumes. Recent advances in deep learning have introduced powerful alternatives, particularly convolutional neural networks (CNNs) and transformer-based architectures, which can learn hierarchical and long-range structural patterns directly from seismic data. This review provides a comprehensive synthesis of studies employing CNNs, multi-scale networks, U-Net variants, attention modules, and early transformer models for automated fault detection. We examine their principles, architectures, datasets, evaluation metrics, strengths, and limitations in comparison to conventional methods. Furthermore, we analyze the role of hybrid CNN–Transformer frameworks and discuss challenges such as limited labeled data, computational constraints, and generalization across geological settings. Finally, we outline future research directions, including semi-supervised learning, domain-informed modeling, and scalable architectures suited for high-resolution 3D seismic interpretation.
References
[1] S. Sukmono et al., “Integration of 3D seismic attributes for preliminary shallow geohazard identification in deep water exploration area with no well data,” First Break, vol. 35, no. 8, 2017.
[2] D. J. Tearpock et al., Applied Three-Dimensional Subsurface Geological Mapping: With Structural Methods. Pearson, 2020.
[3] R. A. Schultz, U. Mutlu, and A. Bere, “Critical issues in subsurface integrity,” in ARMA US Rock Mechanics/Geomechanics Symposium, 2016.
[4] C. L. Liner and T. A. McGilvery, The Art and Science of Seismic Interpretation. Springer, 2019.
[5] H. Di and D. Gao, “Gray-level transformation and Canny edge detection for 3D seismic discontinuity enhancement,” Computers & Geosciences, vol. 72, pp. 192–200, 2014.
[6] W. Xiong et al., “Seismic fault detection with convolutional neural network,” Geophysics, vol. 83, no. 5, pp. O97–O103, 2018
[7] L. Mattéo et al., “Automatic fault mapping in remote optical images and topographic data with deep learning,” J. Geophys. Res. Solid Earth, vol. 126, no. 4, 2021
[8] N. Roy and R. Sarkar, “A review of seismic damage of mountain tunnels and probable failure mechanisms,” Geotechnical and Geological Engineering, vol. 35, no. 1, pp. 1–28, 2017.
[9] A. M. C. Şengör, “Cross-faults and differential stretching of hanging walls in regions of low-angle normal faulting: examples from western Turkey,” Geological Society, London, Special Publications, vol. 28, no. 1, pp. 575–589, 1987.
[10] H. Di et al., “Improving seismic fault detection by super-attribute-based classification,” Interpretation, vol. 7, no. 3, pp. SE251–SE267, 2019.
[11] A. Libak, B. Alaei, and A. Torabi, “Fault visualization and identification in fault seismic attribute volumes: Implications for fault geometric characterization,” Interpretation, vol. 5, no. 2, pp. B1–B16, 2017.
[12] I. Cohen, N. Coult, and A. A. Vassiliou, “Detection and extraction of fault surfaces in 3D seismic data,” Geophysics, vol. 71, no. 4, pp. P21–P27, 2006.
[13] M. Alam et al., “Interpretation of geologic facies in seismic volume using key rock elastic properties and high-definition facies templates,” Interpretation, vol. 5, no. 2, pp. SE11–SE27, 2017.
[14] M. Y. Alfarhan, Concurrent Seismic Events Detection Using Deep Learning, 2020.
[15] B. Zhang et al., “Semiautomated fault interpretation based on seismic attributes,” Interpretation, vol. 2, no. 1, pp. SA11–SA19, 2014.
[16] Q. S. Imran et al., “Automated fault detection and extraction under gas chimneys using hybrid discontinuity attributes,” Applied Sciences, vol. 11, no. 16, p. 7218, 2021.
[17] O. M. Saad et al., “Self-attention deep image prior network for unsupervised 3-D seismic data enhancement,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–14, 2021.
[18] D. F. Fernandez et al., “Machine learning assisted seismic interpretation of top and base carbonate for Earth model building: A case study of carbonate platforms from southern Gulf of Mexico,” in First International Meeting for Applied Geoscience & Energy, SEG, 2021.
[19] Y. Ao et al., “Seismic structural curvature volume extraction with convolutional neural networks,” IEEE Trans. Geosci. Remote Sens., vol. 59, no. 9, pp. 7370–7384, 2020.
[20] D. Wang et al., “Application of multiscale learning neural network based on CNN in bearing fault diagnosis,” J. Signal Process. Syst., vol. 91, no. 10, pp. 1205–1217, 2019.
[21] D. Wang and G. Chen, “Seismic stratum segmentation using an encoder–decoder convolutional neural network,” Mathematical Geosciences, vol. 53, no. 6, pp. 1355–1374, 2021.
[22] F. Gao et al., “SD-CNN: A shallow-deep CNN for improved breast cancer diagnosis,” Computerized Medical Imaging and Graphics, vol. 70, pp. 53–62, 2018.
[23] M. Hahn, “Theoretical limitations of self-attention in neural sequence models,” Trans. Assoc. Comput. Linguistics, vol. 8, pp. 156–171, 2020.
[24] S. Plakias and Y. S. Boutalis, “Fault detection and identification of rolling element bearings with Attentive Dense CNN,” Neurocomputing, vol. 405, pp. 208–217, 2020.
[25] S. H. Elinjulliparambil, “3D reconstruction from monocular videos using neural radiance fields (NeRF),” Int. J. Emerging Res. Eng. Technol., vol. 3, no. 4, pp. 115–127, 2022.
[26] S. H. Elinjulliparambil, “Vision transformers (ViT) for small-scale image classification with token reduction,” Int. J. AI, Big Data, Comput. Manage. Stud., vol. 3, no. 4, pp. 115–122, 2022.
