Research on pattern recognition technology for a highperformance optical fiber vibration sensor

Abstract

In this paper, the optical fiber vibration sensor based on Mach-Zehnder Interference (MZI) principle is designed and researched, which can improve the ability to recognize the physical intrusion events for underground power pipelines by means of pattern recognition technology. Through data collection, analysis and processing, the vibration event database of key intrusion events in typical environment is established. By using coherent optical beat frequency, vibration recognition algorithm, machine learning and other technologies, the frequency domain diagram and time domain diagram of different vibration signals are analyzed and judged. In the end, the feature values of optical fiber vibration sensor such as frequency range, amplitude and waveform density are extracted.