Vehicular traffic density state estimation using Support Vector Machine

Road traffic congestion is a severe problem worldwide due to increased motorization, urbanization and population growth. Traffic congestion reduces the efficiency of the transportation infrastructure of a city; increases travel time, fuel consumption and air pollution, and leads to increased user frustration and fatigue. Reducing traffic congestion can improve traffic flow, reduce travel times and the environmental impact. The main objective of this paper is to consider the problem of vehicular traffic density to determine the low and high traffic conditions. To determine the traffic firstly we determine the texture features. Based on the texture features we determine the various traffic conditions. The procedure includes background subtraction from which we obtain the difference image and we apply the Support Vector Machine (SVM) procedure on a given captured image. Experimental result shows that the approaches are very efficient and produce up to 90% accuracy.