|
|
基于深度学习的等值反磁通瞬变电磁数据反演
|
Abstract:
等值反磁通瞬变电磁法(OCTEM)是一种新兴的地球物理勘探方法。它在城市和工程勘测中越来越受欢迎。本研究采用深度学习(DL)反演方法来解决等值反磁通瞬变电磁反演问题。首先我们创建了一个合成样本数据集，包括50,000个电阻率模型–OCTEM响应对组成的合成样本数据集，用于网络训练。采用流行的卷积神经网络架构U-Net和残差块(Res Net)来构建我们的深度学习反演模型OCTEM InvNet。实验反演实例表明，所提出的OCTEM InvNet能产生准确可靠的反演结果。OCTEM InvNet在重建地下电阻率结构方面表现较好，并能实现瞬时反演。
Opposing-coils transient electromagnetic (OCTEM) is a burgeoning geophysical exploration method. It is gaining popularity in urban and engineering surveys due to the advantages of high efficiency and resolution. In this study, a deep learning (DL) inversion method is proposed to solve the OCTEM inverse problem. We create a synthetic sample dataset consisting of 50,000 resistivity model-OCTEM response pairs for network training and employ the popular convolutional neural network architectures U-Net and residual block (Res Net) to construct our deep learning inversion model OCTEM InvNet. The experimental inversion examples demonstrate that the proposed OCTEM InvNet can produce accurate and reliable inversion results. OCTEM InvNet performs well in reconstructing the subsurface resistivity structure and enables instantaneous inversion.
| [1] | Xi, Z.Z., Long, X., Zhou, S., Huang, L., Song G., Hou, H.T. and Wang, L. (2016) Opposing Coils Transient Electromagnetic Method for Shallow Subsurface Detection. Chinese Journal of Geophysics, 59, 551-559. https://doi.org/10.1002/cjg2.30006 |
| [2] | Wang, Y., Xi, Z.Z., Jiang, H., Hou, H.T., Zhou, S. and Fan, F.L. (2017) The Application Research on the Detection of Karst Disease of Airport Runway Based on OCTEM. Geophysical and Geochemical Exploration, 41, 360-363. |
| [3] | Long, X., Xi, Z.Z., Zhou, S., Hou, H.T., Wang, L. and Xue, J.P. (2020) Detection Capability of Opposing Coils Transient Electromagnetic Method for Thin Layers. Progress in Geophysics, 35, 753-759. |
| [4] | Zhang, X.Y., Li, N.B. and Pei, S.J. (2022) The Application of Opposing Coils Transient Electromagnetics Method in Subway Karst Investigation. Railway Investigation and Surveying, 48, 52-56. |
| [5] | Wang, L., Li, H., Wang, D.H., Zhou, S., Zhang, W., Long, X., Yang, J. and Wang, Q. (2023) Urban Geophysical Exploration: Case Study in Chengdu International Bio-City. Journal of Geophysics and Engineering, 20, 830-840. https://doi.org/10.1093/jge/gxad049 |
| [6] | Fan, T., Xue, G.Q., Li, P., Yan, B., Bao, L., Song, J.Q., Ren, X. and Li, Z.L. (2022) TEM Real-Time Inversion Based on Long-Short Term Memory Network. Chinese Journal of Geophysics, 65, 3650-3663. |
| [7] | Sun, H.F., Zhang, N.Y., Liu, S.B., Li, D.R., Chen, C.D., Ye, Q.Y., Xue, Y.G. and Yang, Y. (2019) L1-Norm Based Nonlinear Inversion of Transient Electromagnetic Data. Chinese Journal of Geophysics, 62, 4860-4873. |
| [8] | Ling, T.H., Qin, J., Song, Q. and Hua, P. (2020) Intelligent Displacement Back-Analysis Based on Improved Particle Swarm Optimization and Neural Network and Its Application. Journal of Railway Science and Engineering, 17, 2181-2190. |
| [9] | Xu, Z.Y., Fu, N.Y., Zhou, J. and Fu, Z.H. (2022) Comparison of Nonlinear Optimization Inversion Algorithms of Transient Electromagnetic Method. Journal of Jilin University (Earth Science Edition), 3, 744-753. |
| [10] | Wang, H., Liu, M.L., Xi, Z.Z., Peng, X.L. and He, H. (2018) Magnetotelluric Inversion Based on BP Neural Network Optimized by Genetic Algorithm. Chinese Journal of Geophysics, 61, 1563-1575. |
| [11] | Wang, H., Yan, J.Y., Fu, G.M. and Wang, X. (2020) Current Status and Application Prospect of Deep Learning in Geophysics. Progress in Geophysics, 35, 642-655. |
| [12] | Yu, S.B., Sheng, Y.H. and Zhang, Y. (2023) CG-DAE: A Noise Suppression Method for Two-Dimensional Transient Electromagnetic Data Based on Deep Learning. Journal of Geophysics and Engineering, 20, 600-609. https://doi.org/10.1093/jge/gxad035 |
| [13] | Davood, M. (2020) One-Dimensional Deep Learning Inversion of Electromagnetic Induction Data Using Convolutional Neural Network. Geophysical Journal International, 222, 247-259. https://doi.org/10.1093/gji/ggaa161 |
| [14] | Vladimir, P. and Andrei, S. (2021) Inversion of 1D Frequency-and Time-Domain Electromagnetic Data with Convolutional Neural Networks. Computers & Geosciences, 149, Article 104681. https://doi.org/10.1016/j.cageo.2020.104681 |
| [15] | Wu, S.H., Huang, Q.H. and Zhao, L. (2021) Convolutional Neural Network Inversion of Airborne Transient Electromagnetic Data. Geophysical Prospecting, 69, 1761-1772. https://doi.org/10.1111/1365-2478.13136 |
| [16] | Liu, W., Wang, H., Xi, Z.Z., Zhang, R.Q. and Huang, X.D. (2022) Physics-Driven Deep Learning Inversion with Application to Magnetotelluric. Remote Sensing, 14, Article 3218. https://doi.org/10.3390/rs14133218 |
| [17] | Wang, Y.Q., Wang, Q., Lu, W.K. and Li, H. (2022) Physics-Constrained Seismic Impedance Inversion Based on Deep Learning. IEEE Geoscience and Remote Sensing Letters, 19, 1-5. https://doi.org/10.1109/LGRS.2021.3072132 |
| [18] | Li, F.D., Guo, Z.W., Pan, X.P., Liu, J., Wang, Y. and Gao, D.W. (2022) Deep Learning with Adaptive Attention for Seismic Velocity Inversion. Remote Sensing, 14, Article 3810. https://doi.org/10.3390/rs14153810 |
| [19] | Liu, Z.G., Chen, H., Ren, Z.Y., Tang, J.T., Xu, Z.M., Chen, Y.P. and Liu, X. (2021) Deep Learning Audio Magnetotellurics Inversion Using Residual-Based Deep Convolution Neural Network. Journal of Applied Geophysics, 188, Article 104309. https://doi.org/10.1016/j.jappgeo.2021.104309 |
| [20] | Wu, X., Xue, G.Q., Zhao, Y., Lv, P.F., Zhou, Z. and Shi, J.J. (2022) A Deep Learning Estimation of the Earth Resistivity Model for the Airborne Transient Electromagnetic Observation. Journal of Geophysical Research: Solid Earth, 127, e2021JB023185. https://doi.org/10.1029/2021JB023185 |
| [21] | Liu, X, Craven, A.J. and Tschirhart, V. (2023) Retrieval of Subsurface Resistivity from Magnetotelluric Data Using a Deep-Learning-Based Inversion Technique, Minerals, 13, Article 461. https://doi.org/10.3390/min13040461 |
| [22] | Wang, L., Long, X., Wang, T.T., Xi, Z.Z., Chen, X.P., Zhong, M.F. and Dong, Z.Q. (2022) Application of the Opposing-Coils Transient Electromagnetic Method in Detection of Urban Shallow Cavities. Geophysical and Geochemical Exploration, 46, 1289-1295. |
| [23] | Ronneberger, O., Fischer, P. and Brox, T. (2015) U-Net: Convolutional Networks for Biomedical Image Segmentation. Medical Image Computing and Computer-Assisted Intervention—MICCAI 2015, Munich, 5-9 October 2015, 234-241. https://doi.org/10.1007/978-3-319-24574-4_28 |
| [24] | He, K.M., Zhang X.Y., Ren S.Q. and Sun, J. (2015) Deep Residual Learning for Image Recognition, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, 27-30 June 2016, 770-778. https://doi.org/10.1109/CVPR.2016.90 |
| [25] | Kingma, D. and Ba, J. (2014) Adam: A Method for Stochastic Optimization. arXiv: 1412.6980. https://doi.org/10.48550/arXiv.1412.6980 |
