Vehicle Type Classification using Vehicle Detector Data from ETC Lane

An Electronic Toll Collection (ETC) lane is equipped with multiple vehicle detectors, each of which has a number of optical sensors arranged in the vertical direction to detect all vehicles without fail. Various devices are controlled by the vehicle detection information of these vehicle detectors as a trigger. The optical sensors arranged in the vertical direction in each vehicle detector provide one-dimensional data at each sampling time when a vehicle passes through an ETC lane. By arranging these data in chronological order, a binary image of the vehicle's side (which we call the vehicle shadow) can be obtained. In this study, we focused on this vehicle shadow data and investigated whether it could be used to classify the vehicle type.