I developed this real-time intelligent transportation system in my master thesis. It measures various traffic parameters, such as number of vehicles, vehicle class (motorcycle, light vehicle, heavy vehicle), speed of vehicle. It also estimates total number of lane changing vehicles – this can be used for traffic congestion prediction. This system automatically warns an operator when speeding vehicle is detected. It correctly works under various wheater conditions (sunny, cloudy, snowing…).
Performance: lane change detection probabilty – 90%, speed estimation error – 2.18 km/h, vehicle counting accuracy – 97.64%, vehicle classification accuracy – 98.66%
Automated road lanes detection and vanishing point estimation is used in camera calibration phase.
System is implemented in Microsoft Visual C++ 2010 using OpenCV library.
Post time: Mar-20-2017