DETECTING AND COUNTING VEHICLES FROM SMALL LOW-COST UAV IMAGES
Cheng Peng-gen, Guoqing Zhou, Zezhong Zheng · 2009
In recent years, many civil users have been interested in unmanned aerial vehicle (UAV) for traffic monitoring and traffic data collection because they have the ability to cover a large area, focus resources on the current problems, travel at higher speeds than ground vehicles, and are not restricted to traveling on the road network. This paper presents a method for detecting and counting vehicles from UAV video flow. The algorithm for vision-based detection and counting of vehicles in monocular image sequences for traffic scenes have been developed. In the algorithm, video frame-to-frame matching to track vehicle is one of important steps. Dynamic vehicles are identified using both background elimination and background registration techniques. The background elimination method uses concept of least squares to compare the accuracies of the current algorithm with the already existing algorithms. The background registration method uses background subtraction which improves the adaptive background mixture model and makes the system learn faster and more accurately, as well as adapt effectively to changing environments. In addition, because of high data sampling rates of video flow, resampling of video flow is also analyzed and discussed. The objective of this research is to monitor activities at traffic intersections for detecting congestions, and then predict the traffic flow.