TRAFFIC FLOW MEASUREMENT USING DOUBLE SLIT IMAGE

This paper proposes a video based one directional traffic flow measurement method using line sensor camera. Line sensor camera is inherently provided with the high resolution and flexible line capturing speed. Line sensor camera outputs pseudo two-dimensional image that consists of space domain and time domain. The research detects occupancy, time headway and time between two cars from information along time axis in a slit image. Vehicles are extracted using fundamental image processing. In double slit configuration, spot speed of vehicles is detected by the time difference of its appearance at each slit. Normalized vehicle images are restored (recovered line sensor output images in the case that a capturing speed of a camera would be synchronized with a vehicle speed). Using line sensor camera in double slit configuration, we estimate vehicle length, space headway and distance between two cars. Traffic volume is counted, rate of flow, occupancy, time mean speed, average time headway and average time between two cars, as macroscopic traffic flow parameters for designated measurement interval. Top-view of normalized vehicle images is represented. The normalized images are available for classification and identification of vehicles.