ON OBTAINING INITIAL APPROXIMATION FOR FULL WAVE INVERSION PROBLEM USING CONVOLUTIONAL NEURAL NETWORK

I. B. Petrov; Петров И. Б.; A. S. Stankevich; Станкевич А. С.; A. V. Vasyukov; Васюков А. В.

doi:10.31857/S2686954322600732

ON OBTAINING INITIAL APPROXIMATION FOR FULL WAVE INVERSION PROBLEM USING CONVOLUTIONAL NEURAL NETWORK

Autores: Petrov I.¹, Stankevich A.¹, Vasyukov A.¹
Afiliações:
1. Moscow Institute of Physics and Technology (National Research University)
Edição: Volume 512, Nº 1 (2023)
Páginas: 58-64
Seção: МАТЕМАТИКА
URL: https://journals.rcsi.science/2686-9543/article/view/139277
DOI: https://doi.org/10.31857/S2686954322600732
EDN: https://elibrary.ru/SWNWFX
ID: 139277

Citar

Texto integral

Acesso aberto
Acesso é fechado

Acesso está concedido
Acesso é fechado

Somente assinantes

Resumo
Sobre autores
Bibliografia
Arquivos suplementares
Estatísticas

Resumo

The paper considers the problem of choosing the initial approximation when using gradient optimization methods for solving the inverse problem of restoring the distribution of velocities in a heterogeneous continuous medium. A system of acoustic equations is used to describe the behavior of the medium, and a finite-difference scheme is used to solve the direct problem. L-BFGS-B is used as a gradient optimization method. Adjoint state method is used to calculate the gradient of the error functional with respect to the medium parameters. The initial approximation for the gradient method is obtained using a convolutional neural network. The network is trained to predict the distribution of velocities in the medium from the wave response from it. The paper shows that a neural network trained on responses from simple layered structures can be successfully used to solve the inverse problem for a complex Marmousi model.

Palavras-chave

acoustic inversion, numerical optimization, adjoint state method, machine learning, deep learning, convolutional neural networks

Bibliografia

Tarantola A. Inversion of seismic reflection data in the acoustic approximation // Geophysics. 1984. V. 49. № 8. P. 1259–1266.
Ovcharenko O., Kazei V., Peter D., Alkhalifah T. Variance-based model interpolation for improved full-waveform inversion in the presence of salt bodies // Geophysics. 2018. V. 83. № 5. P. R541–R551.
Sun H., Demanet L. Extrapolated full-waveform inversion with deep learning EFWI-CNN // Geophysics. 2020. V. 85. № 3. P. R275–R288.
Li H., Schwab J., Antholzer S., Haltmeier M. NETT: solving inverse problems with deep neural networks // Inverse Problems. 2020. V. 36. № 6. P. 065005.
Kothari K., de Hoop M., Dokmani’c I. Learning the Geometry of Wave-Based Imaging // Advances in Neural Information Processing Systems. 2020. V. 33. P. 8318–8329.
Gahlmann T., Tassin P. Deep neural networks for the prediction of the optical properties and the free-form inverse design of metamaterials // Phys. Rev. B. 2022. V. 106. № 8. P. 085408.
Adler A., Araya-Polo M., Poggio T. Deep Learning for Seismic Inverse Problems: Toward the Acceleration of Geophysical Analysis Workflows // IEEE Signal Processing Magazine. 2021. V. 38. № 2. P. 89–119.
Yang F., Ma J. Deep-learning inversion: a next generation seismic velocity-model building method // Geophysics. 2019. V. 84. № 4. P. R583–R599.
Mast T.D., Hinkelman L.M., Metlay L.A., Orr M.J., Waag R.C. Simulation of ultrasonic pulse propagation, distortion, and attenuation in the human chest wall // Journal of the Acoustical Society of America. 1999. V. 6. P. 3665–3677.
Golubev V., Shevchenko A., Khokhlov N., Petrov I., Malovichko M. Characteristic Scheme for the Acoustic System with the Piece-Wise Constant Coefficients // International Journal of Applied Mechanics. 2022. V. 14. № 2. P. 2250002.
Levander A.R. Fourth-order finite-difference P-SV seismograms // Geophysics. 1988. V. 53. № 11. P. 1425–1436.
Martin R., Komatitsch D., Ezziani A. An unsplit convolutional perfectly matched layer improved at grazing incidence for seismic wave propagation in poroelastic media // Geophysics. 2008. V. 73. № 4. P. T51–T61.
Paszke A., Gross S., Massa F., Lerer A., Bradbury J., Chanan G., Killeen T., Lin Z., Gimelshein N., Antiga L., Desmaison A., Kopf A., Yang E., DeVito Z., Raison M., Tejani A., Chilamkurthy S., Steiner B., Fang L., Bai J., Chintala S. PyTorch: An Imperative Style, High-Performance Deep Learning Library // Advances in Neural Information Processing Systems 32. 2019. P. 8024–8035.
Li D., Xu K., Harris J.M., Darve E. Coupled Time-lapse Full Waveform Inversion for Subsurface Flow Problems using Intrusive Automatic Differentiation // 2019. arXiv: 1912.07552.
Xu K., Li D., Darve E., Harris J.M. Learning Hidden Dynamics using Intelligent Automatic Differentiation // 2019. arXiv: 1912.07547.
Byrd R.H., Nocedal J., Schnabel R.B. Representations of quasi-Newton matrices and their use in limited memory methods // Mathematical Programming. 1994. V. 63. № 1. P. 129–156.
Plessix R.-E. A review of the adjoint-state method for computing the gradient of a functional with geophysical applications // Geophysical Journal International. 2006. V. 167. № 2. P. 495–503.
Ronneberger O., Fischer P., Brox T. U-Net: Convolutional Networks for Biomedical Image Segmentation // CoRR. 2015. V. abs/1505.04597. arXiv: 1505.04597.
Vasyukov A.V., Nikitin I.S., Stankevich A.S., Golubev V.I. Deep convolutional neural networks in Seismic Exploration problems // Interfacial Phenomena and Heat Transfer. 2022. V. 10. № 3. P. 61–74.
Brougois A., Bourget M., Lailly P., Poulet M., Ricarte P., Versteeg R. Marmousi, model and data // EAEG Workshop – Practical Aspects of Seismic Data Inversion. 1990.