Improve the negative effects of unknown disturbance on the event recognition for Phi-OTDR

Abstract

This paper presents a new method to reduce the high false alarm rate. When the Phi-OTDR system is applied in complex environment, irrelevant event signals similar to the target event often interfere with the recognition of the target event and leads to false positive in the system. This method includes two key components: one is the automatic labeling mechanism of unknown information, which is used to find unknown events easily confused with target events in real environment; the other is the improved Region of Interest (ROI) module based on contrast clustering and EBMs, which can detect target event signals in real complex environment and reduce the negative impact of unknown events on target event recognition.