Biologically inspired composite vision system for traffic monitoring
Han Ming Show, Lin Lin, Ramesh Bharath, Xiang Cheng · 2016
We present a novel vision-based traffic monitoring system, which is inspired by the visual structure found in raptors, for tracking expressway vehicles and estimation of their real-world speed over a longer range compared to the typical, single-camera monitoring systems. This vision system features a license plate detection camera that is triggered whenever there is an instance of overspeeding. The proposed system provides accurate speed information for overspeed vehicle detection by estimating the real-world coordinates of the vehicle. To this end, a speed calculation algorithm is designed for the composite vision information acquired by the system, and with the aid of a license plate detection camera, identity information of the over-speeding vehicles can be recorded for law enforcement. The calculated speed was verified using RADAR speed detection systems and smartphone applications, and the deviation was found to be within ±3 km/hr compared to the actual driving speed. In summary, the proposed system provides a novel solution to improve the capabilities of traffic monitoring using vision based methods.