Fast and Storage-Optimized Compressed Domain Vibration Detection and Classification for Distributed Acoustic Sensing

Distributed acoustic sensing (DAS) is an attractive technology that can turn existing fibre optic networks to real-time distributed vibration sensors. However, it encounters great challenges in handling the transmission, storage, and processing of voluminous data streams which are orders of magnitude larger than that collected from point sensors. The gap between intensive data generated by φ-OTDR and modern computing system with limited reading/writing speed and storage capacity leads to restrictions on many applications. Compressive sensing (CS) is a revolutionary signal acquisition method that allows a signal to be acquired and reconstructed with significantly fewer samples than that required by Nyquist-Shannon theorem. The data size is greatly reduced in the sampling stage, but the reconstruction of the compressed data is still time and computation consuming. To address this challenge, we propose to map the feature extractor from Nyquist-domain to compressed-domain and therefore vibration detection and classification can be directly implemented in compressed-domain. The measured results show that our method can be used to reduce the transmitted data size by 70% while achieves a true positive rate (TPR) of 99.4% and a false positive rate (FPR) of 0.04% of positioning along 5-km-long sensing fibre and a classification accuracy of 95.05% on a 5-class classification task.