Attention based Temporal convolutional network for ϕ-OTDR event classification
Manling Tian, Hui Dong, Kuanglu Yu
Abstract
We designed a new attention based temporal convolutional network combined with bidirectional long short term memory model named ATCN-BiLSTM for Φ-OTDR event classification, achieving average classification accuracy of 99.6% on three types of events.
Topics & Concepts
Optical time-domain reflectometerComputer scienceEvent (particle physics)Artificial intelligenceConvolutional neural networkPattern recognition (psychology)Optical fiberTelecommunicationsFiber optic splitterFiber optic sensorPhysicsQuantum mechanicsTime Series Analysis and ForecastingAdvanced Fiber Optic SensorsNeural Networks and Reservoir Computing