Abstract
There is a high false alarm rate in identifying event process of traditional long-distance optical fiber pre-warning system. To solve the problem, an optical fiber pre-warning system was proposed based on an improved neural network and its adaptability to different environments was studied. Firstly, Phi-OTDR technology was applied to design the distributed sensing part of the system. And then, an improved neural network was used for the signal recognition part to identify and classify intrusion events. Finally, experiments were carried out in three cases to analyze the adaptability of the system. The results show that the system can provide an excellent classification effect in the recognition of optical fiber vibration signals, and it also has better adaptability under different environmental conditions.
Translated title of the contribution | An Improved Neural Network Based Research on Generalization of an Optical Fiber Pre-Warning System |
---|---|
Original language | Chinese (Traditional) |
Pages (from-to) | 649-657 |
Number of pages | 9 |
Journal | Beijing Ligong Daxue Xuebao/Transaction of Beijing Institute of Technology |
Volume | 41 |
Issue number | 6 |
DOIs | |
Publication status | Published - Jun 2021 |