Abstract

A smart distributed acoustics/vibration sensing (DAVS) scheme with dual path network is proposed for complex applications in this manuscript. The dual path network has high parameter efficiency and strong generalization ability, which provides the reliability and the robustness. Novel spatial short-time frequency spectrum datasets are constructed and the distributed spatial features of disturbances are well utilized. To the best of our knowledge, it is the first time that such a hyper-deep neural network is applied into distributed fiber sensing field and the unique spatial dimension of DAVS is well exploited in datasets. Moreover, the scheme presented satisfying event-classification results in field tests of complicated railway safety monitoring as proof-of-concept. It provides a solution to high accuracy multi-event classification in complex and time-varying distributed safety monitoring applications.

© 2018 The Author(s)