Limit Line Of Vehicles Based On The Provisions Of Sections Of The Video Image Detection Technology

Traffic Laws have strict provisions of dedicated lanes.No motor vehicles but allowable cars are allowed to drive in the dedicated lanes.The dedicated lanes are usually be monitored by employing people who watching the roads all the day.This method wastes time and money,so it is inefficient for the traffic management.Based on research of video detecting and feature analyze,this paper design and implement a illegal vehicles detecting system,which detect the illegal vehicles automatically.Two problems are involved to build this system:extracting feature of vehicles and developing classifier for vehicles.Aiming to the requirement of describing the feature of vehicles exactly,Haar feature rectangle which based on statistic theory is involved to extract feature of vehicles,which can code the area information of picture,and process the picture in Haar space.Aiming to detect the vehicles in real time,AdaBoost arithmetic is used to develop classifier for vehicles.The weak classifiers,which related to feature of illegal vehicles,are trained,and the best of them are chosen to build strong classifier.A cascade method is employed to combine the strong classifiers to a cascade classifier,which accelerates the detection speed significantlyA system is showed in the end of this paper,which employs the real traffic video of Qingdao to detect the illegal vehicles.Trucks are chosen as example of illegal vehicles. The experiment shows that most of the illegal vehicles could be detected by this system.