Abstract
This paper proposes a real-time multi-class disturbance detection algorithm based on YOLO for distributed fiber vibration sensing. The algorithm achieves real-time detection of event location and classification on external intrusions sensed by distributed optical fiber sensing system (DOFS) based on phase-sensitive optical time-domain reflectometry (Φ-OTDR). We conducted data collection under perimeter security scenarios and acquired five types of events with a total of 5787 samples. The data is used as a spatial–temporal sensing image in the training of our proposed YOLO-based model (You Only Look Once-based method). Our scheme uses the Darknet53 network to simplify the traditional two-step object detection into a one-step process, using one network structure for both event localization and classification, thus improving the detection speed to achieve real-time operation. Compared with the traditional Fast-RCNN (Fast Region-CNN) and Faster-RCNN (Faster Region-CNN) algorithms, our scheme can achieve 22.83 frames per second (FPS) while maintaining high accuracy (96.14%), which is 44.90 times faster than Fast-RCNN and 3.79 times faster than Faster-RCNN. It achieves real-time operation for locating and classifying intrusion events with continuously recorded sensing data. Experimental results have demonstrated that this scheme provides a solution to real-time, multi-class external intrusion events detection and classification for the Φ-OTDR-based DOFS in practical applications.
Highlights
The distributed optical fiber sensing system (DOFS) system can locate the position of the disturbance in the spatial domain and acquire the vibration information of the disturbance in the temporal domain
This paper proposes a real-time multi-class disturbance detection method based on YOLO algorithm for Φ-OTDR
Using the YOLO algorithm based on Darknet53 and Feature Pyramid Networks (FPN), real-time monitoring can be performed on spatial–temporal sensing data acquired from the Φ-OTDR system
Summary
After being proposed in 2005 [1], the phase-sensitive optical time-domain reflection technique has been widely used [2] in geological exploration [3–5], partial discharge monitoring [6,7], traffic sensing [8,9], marine health monitoring [10], and perimeter security [11,12], etc. A lot of works are proposed based on traditional classification algorithms that use human-extracted signal features for learning to classify disturbances [15]. Sun et al artificially extracts multiple features and perform correlation analysis for dimensionality reduction on the disturbance signals in the spatial–temporal domain and use three RVM classifiers to classify the three types of intrusions, achieving an accuracy rate of 97.8% [17]. These traditional classification algorithms belong to “expert systems”, and they require human-determined features. Different from their work, we demonstrate a new method for detecting and classifying multi-class disturbance events and achieving real-time operation, making it more suitable for practical applications. A real-time multi-class disturbance detection scheme for Φ-OTDR-based DOFS is provided to the community, which we believe will have a positive effect on practical applications, especially online monitoring scenarios
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
