The Study Of Static And Dynamic Traffic Information Processing And Application Based On The Unmanned Aerial Vehicle

With the rising number of vehicles,the urban traffic is facing a lot of pressure.Traffic congestion is very serious,it has a impact on the development of social economy and people’s daily life.Reducing the traffic pressure effectively and easing traffic congestion have become important research subjects for the relevant departments such as traffic planning department,transportation design department and traffic management department.While to obtain traffic data is the first step to deal with these traffic problems.Aims at the static and dynamic traffic information needed in traffic simulation modeling,this paper use a camera equipped on the UAV to collect traffic images and videos,from which the static and dynamic information can be obtained,based on the main trunk road areas in Nanshan district of Shenzhen city.In the static information process,on the basis of the existing image motion blur handling method,the image for motion blur will be analyzed,then the methods of extracting the motion blurred direction and the length is improved.using the existing image restoration methods this paper achieves recovery of aerial images.To achieve road marking detection,the maximum inter-class variance will be used.In terms of dynamic information process,combining with the hybrid gaussian background modeling Method the paper realizes the moving object in aerial video.and using mean-shift method and kalman filter realizes single objective and multiple target tracking respectively in moving target tracking.In order to verify the feasibility of the improved method,simulation experiments were carried out using simulated blurred images,the results show that when the blurred length is greater than 10 pixels,the paper’ s methods improve the accuracy of blurred direction and blurred length,after that the paper applies them to real aerial images.Combining with the existing methods the detection accuracy of road marking will be improved.Finally the author extracts complete moving targets.In this the paper will provide some data base for traffic simulation modeling and other traffic analysis.