Preliminary Research on Vehicle Speed Detection using Traffic Cameras
Chen Wang, Aibek Musaev · 2019
Recently, research on detection of abnormal vehicle behavior is getting increasing attention through various approaches, including interval speed measurement using radars, binocular surveillance cameras, and infrared speed sensors. In most areas in the U.S., such detection is still performed manually by the highway patrol. This paper presents a novel approach for detecting abnormal vehicle behavior using the traditional traffic surveillance cameras. Once the vehicle is within the camera field of view (FOV), its estimated speed is automatically detected based on designated references with respect to dashed white lines on roads, such that the speed of the vehicle is computed using techniques. The proposed method, if validated and facilitated, can largely reduce the cost of upgrading surveillance infrastructure as well as the extra human administrative burden.