A combined edge detection and background differencing image processing approach for real-time traffic analysis
Mohamed Fathy, Mohammed Yakoob Siyal · Road and transport research · 1995
Road traffic data collection and analysis is very important for both designing new roads and maintaining the existing road traffic networks. Vision based vehicle detection techniques have been investigated by various researchers for a number of years and some progress have been made in this regard. In recent years, edge detection techniques have been used for real-time vehicle detection. However, edge detectors only extract the edges of a scene irrespective of whether it belongs to the background details of the objects themselves. Therefore, to separate these components of the scene, extra information is required. In this paper, a new vehicle detection method is described. This method combines background differencing and edge detection techniques to separate objects from their backgrounds. The method works well under various lighting and weather conditions. The new vehicle detection technique can be used to determine other road traffic parameters such as the number of vehicles, and types of vehicles. This method of vehicle detection works in real -time on a low-cost 80386 based microcomputer system.