|
Today 's hot research topic in intelligent transportation systems , sports vehicles unattended , it relates to the various aspects of computer vision, information collection , video image processing technology . The effective detection and accurate tracking of the movement of vehicles is a core part of the modern intelligent transportation research . This paper some research, given the movement of vehicles in the context of natural road automatic detection and tracking methods , and the development of software systems , the initial realization of the automatic tracking of the movement of vehicles . Frame difference method - based detection of moving vehicles , and image smoothing , enhancement, edge detection, noise removal , do analysis and research . Connected region , a new multi- sport target detection notation Highway Traffic moving vehicles in the effective detection and extraction . Vehicle detection based on the match - correction - forecast cycle strategy research vehicle tracking . In two adjacent frames , the moving distance according to the movement of vehicles , so that the characteristics of the change in the area of the image is not large , the moving vehicle and the center of the area of the two parameters to be matched in the adjacent figure . Create motion model using Kalman filtering theory , the prediction by vehicle characteristics such as the center position and size , the goal of the moving object obtained by the prediction and the current frame match , find the corresponding relationship of the moving object in each frame in the image sequence , OK its trajectory , further traffic speed , traffic flow and other important parameters . The experimental results show that , to reduce the search area of the template matching , and to reduce the probability of false match , effectively track capable of moving vehicles . Further, since the algorithm processing parameters , the small amount of data can be real-time tracking of the position of the vehicle in the image . Is designed and implemented using Visual C video moving target detection experimental system , given the framework of the system structure and implementation steps . Finally, analysis of influencing factors of traffic detection accuracy , and propose some solutions .
|