Roberts Kadiķis, Kārlis Freivalds “Efficient Video Processing Method for Traffic Monitoring Combining Motion Detection and Background Subtraction” Proceedings of the Fourth International Conference on Signal and Image Processing 2012 (ICSIP 2012) Lecture Notes in Electrical Engineering Volume 221, 2013, pp 131-141
A computationally efficient video processing algorithm for vehicle detection is presented. The algorithm detects vehicles that arrive on a defined detection line. Inter-frame difference is used to detect the presence of a vehicle and background subtraction is used to find the shape of the object that does not belong to an empty road. If movement is detected, intervals corresponding to the moving objects are created on the detection line. Further processing of intervals allows detecting and counting the objects. The accuracy of the proposed method is analyzed on different roads and in different weather conditions.