Data augmentation and its application in distributed acoustic sensing data denoising
- 24 August 2021
- journal article
- research article
- Published by Oxford University Press (OUP) in Geophysical Journal International
- Vol. 228 (1), 119-133
- https://doi.org/10.1093/gji/ggab345
Abstract
As a data-driven approach, the performance of deep learning models depends largely on the quantity and quality of the training datasets, which greatly limits the application of deep learning to tasks with small datasets. Unfortunately, sometimes we need to use limited small datasets to complete our tasks, such as DAS data denoising. However, using a small dataset to train the network may cause over-fitting, resulting in poor network generalization. To solve this problem, we propose an approach based on the combination of a generative adversarial network and a deep convolutional neural network. First, we used a small noise dataset to train a generative adversarial network to generate synthetic noise samples, and then used these synthetic noise samples to augment the noise dataset. Next, we used the augmented noise dataset and the signal dataset obtained through forward modeling to construct a synthetic training set. Finally, a denoising network based on a convolutional neural network was trained on the constructed synthetic training set. Experimental results show that the augmented dataset can effectively improve the denoising performance and generalization ability of the network, and the denoising network trained on the augmented dataset can more effectively reduce various kinds of noise in the DAS data.Funding Information
- National Natural Science Foundation of China (41974143, 41730422)
This publication has 30 references indexed in Scilit:
- Walkaway VSP using multimode optical fibers in a hybrid wirelineGeophysics, 2016
- ImageNet Large Scale Visual Recognition ChallengeInternational Journal of Computer Vision, 2015
- Vertical seismic optical profiling on wireline logging cableGeophysical Prospecting, 2014
- Distributed acoustic sensing for reservoir monitoring with vertical seismic profilingGeophysical Prospecting, 2014
- Distributed Acoustic Sensing – a new tool for seismic applicationsFirst Break, 2014
- Distributed acoustic sensing for reservoir monitoring with VSPGeophysics, 2013
- Field testing of fiber-optic distributed acoustic sensing (DAS) for subsurface seismic monitoringGeophysics, 2013
- Seismic data acquisition—The new millenniumGeophysics, 2001
- Gradient-based learning applied to document recognitionProceedings of the IEEE, 1998
- Learning long-term dependencies with gradient descent is difficultIEEE Transactions on Neural Networks, 1994