Font Size: a A A

Research On Night Vehicle Recognition And Tracking Based On Video

Posted on:2021-04-01Degree:MasterType:Thesis
Country:ChinaCandidate:Y X WenFull Text:PDF
GTID:2392330629950183Subject:Carrier Engineering
Abstract/Summary:PDF Full Text Request
With the development of social economy and the increase of traffic demand,due to the shortage of traffic supply and the restriction of natural resources,the problems of traffic congestion,traffic safety and environmental pollution have appeared in urban traffic.The solution of these problems can not be achieved by increasing traffic supply and traffic restriction,but must be solved by omnibearing and intelligent traffic control.Traffic data acquisition is necessary for intelligent traffic control.Vehicle identification is the basis of many traffic data,such as traffic flow and violation information,so vehicle detection and tracking is the precondition of intelligent traffic control.It is difficult to detect and track the vehicles at night because of insufficient illumination,difficult to obtain the structure and color features of the vehicles,and complicated road environment.This paper focuses on the night road,vehicle characteristics,vehicle detection and tracking of the following work:1)To study the characteristics of roads at night,the local intensity of the images will change dramatically due to the influence of incoming traffic,and the color will also vary greatly due to the different intensity of the headlights,on the basis of statistical analysis,this paper presents a method of image processing based on statistical frame difference,which can effectively eliminate the local lighting and color mutation,and lay a foundation for subsequent background modeling.2)In the case of reflecting light on the road at night,this paper uses gauss’ s mixed background model to build the road’s night background model,and then extracts the moving vehicle by background subtraction.3)To study the characteristics of vehicle at night,a new method based on the color and shape of taillight is proposed,which can effectively identify the color of taillight by using the joint difference of R,G and B in RGB space.4)Based on the extracted vehicle interval,the mean-shift tracking method is used to track the vehicle and realize the recognition of the vehicle trajectory.5)In order to improve the running efficiency of the system and realize the real-time monitoring of the system,the concept of virtual coil is introduced.
Keywords/Search Tags:Intelligent Traffic, night road feature, night vehicle feature, background model, color feature, virtual coil
PDF Full Text Request
Related items