DAS-N2N: Machine learning Distributed Acoustic Sensing (DAS) signal denoising without clean data

Sacha Lapins,Antony Butcher,J.-Michael Kendall,Thomas S. Hudson,Anna L. Stork,Maximilian J. Werner,Jemma Gunning,Alex M. Brisbourne

DOI: https://doi.org/10.1093/gji/ggad460

2023-11-24

Abstract:This article presents a weakly supervised machine learning method, which we call DAS-N2N, for suppressing strong random noise in distributed acoustic sensing (DAS) recordings. DAS-N2N requires no manually produced labels (i.e., pre-determined examples of clean event signals or sections of noise) for training and aims to map random noise processes to a chosen summary statistic, such as the distribution mean, median or mode, whilst retaining the true underlying signal. This is achieved by splicing (joining together) two fibres hosted within a single optical cable, recording two noisy copies of the same underlying signal corrupted by different independent realizations of random observational noise. A deep learning model can then be trained using only these two noisy copies of the data to produce a near fully-denoised copy. Once the model is trained, only noisy data from a single fibre is required. Using a dataset from a DAS array deployed on the surface of the Rutford Ice Stream in Antarctica, we demonstrate that DAS-N2N greatly suppresses incoherent noise and enhances the signal-to-noise ratios (SNR) of natural microseismic icequake events. We further show that this approach is inherently more efficient and effective than standard stop/pass band and white noise (e.g., Wiener) filtering routines, as well as a comparable self-supervised learning method based on masking individual DAS channels. Our preferred model for this task is lightweight, processing 30 seconds of data recorded at a sampling frequency of 1000 Hz over 985 channels (approx. 1 km of fiber) in $<$1 s. Due to the high noise levels in DAS recordings, efficient data-driven denoising methods, such as DAS-N2N, will prove essential to time-critical DAS earthquake detection, particularly in the case of microseismic monitoring.

Geophysics,Machine Learning,Sound,Audio and Speech Processing

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to remove strong random noise in Distributed Acoustic Sensing (DAS) recordings without using clean data as training samples. Specifically, the paper introduces a weakly - supervised machine - learning method called DAS - N2N, which aims to train a deep - learning model to generate a nearly fully - denoised data copy by using only two copies of the same underlying signal randomly observed and noise - contaminated in two different independent implementations. This method is especially suitable for DAS recordings with high noise levels, can effectively improve the Signal - to - Noise Ratio (SNR) of natural microseismic and icequake events, and shows important value in time - critical DAS seismic detection, especially in microseismic monitoring.

The key to the paper lies in utilizing the characteristics of two optical fibers inside the optical fiber. By splicing (connecting) these two optical fibers at one end of the optical fiber, the light pulses emitted from the DAS interrogation unit can "travel back and forth" along the length of the optical fiber, recording two copies of the same underlying signal with different random measurement noises. In this way, a deep - learning model can be trained to produce an almost fully - denoised data copy using only these noisy data copies. This not only improves the quality of the data but also shows the superiority of DAS - N2N in terms of efficiency and effectiveness compared to standard band - pass/band - stop filtering and white - noise (such as Wiener filtering) filtering procedures and self - supervised learning methods based on masking a single DAS channel.

In conclusion, the main objective of the paper is to propose and verify a new and efficient DAS data - denoising method to address the high - noise - level problem prevalent in DAS recordings, thereby enhancing the accuracy and reliability of seismic monitoring, especially microseismic monitoring.