Vehicle Monitoring System Image Detection Algorithm

Digital image processing technology and monitoring system, the combination of applications is a hot research topic in recent years. Particularly intelligent video technology and traffic monitoring system for the integration of the system have not only a digital video compression, transmission, storage and playback capability, allowing the system to have the image analysis, intelligent monitoring of the characteristics, thus to facilitate the realization of vehicles Segmentation and testing, measurement of traffic-related parameters, such as vehicular traffic movement and direction of statistics, according to the relevant mathematical model, but also to achieve the motion detection vehicles, vehicle tracking and vehicle identification.Based on this subject from the vehicle intelligent video detection system of pre-research project to study through a movement of vehicles used in monitoring the image detection algorithm, for a fixed camera in the video images to identify and track moving targets, Video analysis can be made suitable for integration into compressed video image acquisition front-end module, the use of algorithms to improve the efficiency model to achieve goals and maximize the accuracy of judgement.Video image analysis, background extract accurate or not directly related to the quality of segmentation. Based on this paper were passed on static and dynamic background of the formation of the average method, multi-frame in the value of filtering algorithms and adaptive filtering algorithms a lot of theoretical research and experimental demonstration contrast, made applicable to the movement of vehicles from the background detection system Law, combined with a fixed camera movement of vehicles under the detection of image feature, based on the use of mathematical algorithm separated Morphology of the movement of vehicles from the object. Monitoring for image noise problems, according to video images of vehicles for the campaign of using Kalman filter algorithm moving target tracking and noise filters. And by improving the brink of growth based on the active contour model algorithm to further outline precise targets, improve the quality of segmentation, for the identification model to prepare for. Integrated in these sports image detection algorithm, summed up a new target tracking excellent programme to enable it to effectively overcome the split interference, and the segmentation of the general computing algorithm flaws. Finally the use of the programme on the movement of vehicles video images were monitored simulation. The simulation results show that test results can basically meet the requirements of the moving target detection.