Abstract:Distributed optical sensing (DAS) has the advantages of continuous distribution information retrieval, less environmental impact and low cost. In this paper, a method of vehicle trajectory extraction using DAS system was proposed, which provides the basis for the realization of whole time-space traffic state detection in smart highway environments. First, the free fiber core in the highway communication fiber and the DAS detection equipment based on Φ-OTDR were used to collect the ground vibration signal for the driving vehicles in real-time. Then, the vibration signal was transformed into time-frequency domain channel by channel by S transform to accumulate the energy in the frequency interval where the vehicle vibration signal was located, and the vibration energy of all channels was accumulated to form a spatiotemporal two-dimensional vibration energy response. Finally, the multiple enhanced vehicle trajectories were transformed into multiple singular points in the range-velocity domain by Radon transform, so the vehicle trajectories are separated and extracted unsupervised by extracting singular points in the range-velocity domain. The proposed vehicle trajectory extraction method was verified and tested in the real expressway environment of two-way three lanes. The experimental results showed that compared with the method based on wavelet, the vehicle trajectory is more obvious after being processed by this S transform based algorithm. Meanwhile, compared with the method based on Hough transform, more local peaks is accumulated and clustered in Radon domain. Therefore, it is easier to separate multi-vehicle trajectories with the proposed method.