Compressed-domain Data Classification for Distributed Acoustic Sensing System

One of the challenges in distributed acoustic sensing (DAS) system is how to transmit, store, and process massive streams of data. In this work, we propose a novel classification method in compressed-domain for DAS system Our method compresses the signal to reduce the data size and classify the different vibration events using compressed signal directly. We experimentally verified that 90% classification accuracy can be achieved on a four-class classification task with only 30% of original data size. Classification accuracy under different SNRs and measurement ratios are also investigated in this work.