Research On Vehicle Identification Techniques For Moving Vehicles In Video Sequences

? With the continuous deepening of reform and opening up to our daily lives, China's national economy has been steadily improving, ongoing rectification of urban roads, the expansion which we called for the urgent need for a more efficient traffic management and more comprehensive management. In order to fully monitor, control and management traffic on the need to develop a series of laws and regulations, staff, hardware and software together combination of traffic management system. Traffic management system is the information collection, analysis and processing process, and effective traffic control and management of the premise. The use of advanced signal detection means to obtain information on traffic conditions, to achieve the entire transport system to maximize the rate of using transportation and management of information.Traffic video, management, treatment has increasingly become the vehicle for traffic management, an important means of monitoring in the process of traffic management. The video image of the movement has a very wide range of applications in real time precise positioning of the target vehicle, target identification, classification and tracking in traffic management system. At the same time, video tracking of moving vehicles is a necessary prerequisite in the vehicle identification, traffic control and vehicle counting, and also is the core of intelligent management system. The movement of the vehicle detection and tracking is a very important area of research in the traffic. Tracking process is carried out in the video image the process of object matching, however, due to many factors make the detection of the target vehicle becomes very complex, such as Dynamic changes of the background image, interfering factors of the weather and lighting effects. Of course, many scholars and experts to make this a great deal of research, and also put forward a number of image processing method. Image processing procedure, we should capture the video image, after the background of pretreatment, and then identify the target. There are two main problems, first is the changing of the vehicle size in the vision, second is the intersection of different vehicles will have a great disparity in the images, and resulting in difficulties in matching vehicles. On how to more effectively target matching vehicles after image processing is the main point in this paper. Vehicle matching algorithm based on SIFT technology can greatly improve the accuracy of matching under the case of ensure accurate test results, and elimination of environmental factors on the process of matching vehicles.