Anomaly detectıon in large-scale time-series data acquired by Fiber Optic Distributed Acoustic Sensors (DAS) used for perimeter security and pipeline monitoring is a critical problem in machine learning. However, because of the vast amounts of data to process, it can be time and energy intensive. This study looks at how to reduce detection time and computing costs for this use case. In order to distinguish the acoustic event of interest from the noise and establish a binary detection threshold, we employ a Maximum Eigenvalue Detection (MED) approach in conjunction with a Random Matrix Theory (RMT) precept, namely the Tracy-Widom limit. A pipeline of signal processing techniques is used to assist the algorithm, beginning with applying a Moving Average (MA) filter to remove amplitude swings on the signal, which is represented by a data matrix, and then subsampling it to obtain uncorrelated signals among the subsequent columns to reduce the number of data processed. As a result, we can detect events of interest in less time. Following that, low-pass filtering is employed to eliminate low-frequency coefficients induced by various sorts of environmental and seismic events. Following normalization, the MED method is used to each of the Wishart matrices, which are generated by segmenting the data stream into equal small sub-matrices. RMT is used to set a threshold with a false alarm rate of 0.01 (FAR). The data columns matching to the selected MED values are then injected into a Convolutions Neural Network (CNN) to capture and detect the event of interest. When compared to using solely the CNN, the optimal results from our approach, MED followed by a CNN anomaly detection process, demonstrate a faster detection rate for events in security application.
One of the most prominent applications of fiber optic Distributed Acoustic Sensing (DAS) is Perimeter Security via fence monitoring, which is possible when we attach a fiber to the fence. In this study, we aim to detect and classify events occurring around said fence, such as climbing, cutting, and bending. For this, we investigate Deep Learning algorithms, more specifically Convolutional Neural Networks (CNN), as a mean to detect anomalies and classify them. We recorded 48,445 samples of the mentioned events, which were carefully processed and labeled. From each record, we exploited multiple data instances, resulting in a large enough training dataset to produce a robust classifier. We report the optimum network architecture that suited our study for both the anomaly detection and classification task. The optimal model is tested before and after deployment on-site, we report the quantified performance on a test set via a confusion matrix, and observations about the model’s behaviour on the field. Furthermore, we compare our trials and results on two types of fences, namely rigid and loose, to show how it affects the performance of the trained CNN models, as the signal propagates differently between rigid and loose clotures. We report an overall accuracy of 96.15% for the optimal anomaly detection model, and a lower 52.9% for the 3-class classification model. These results are explained and commented on. Finally, we conclude by providing an educated proposal for future improvements.
In this paper, a fiber optic based acoustic sensing system (FOTAS) is presented. Utilizing such cables as hundreds of acoustic movement detectors has become a novel approach for many security applications. Developed for securing critical infrastructures via utilizing both buried and fence-mounted cables, phase OTDR system FOTAS’s requirements, design, development, test and analysis modules are outlined. The detected movement type, location, magnitude are conveyed to the user via a user-friendly interface. Associated hardware, software, and utilized fiber optic network characteristics are presented. Detailed information about developed novel signal processing and artificial intelligencebased movement detection studies are provided. Real world data indicate that preliminary threat detection performance is more than 90%.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.