Moving Pedestrian Tracking Method based on Mmonocular Vision Technology
Abstract
This paper studied the accurate detection, extraction and tracking of moving objects in video surveillance. The simple main tool kalman filter, the motion detection method and how to predict the next behavior of the movement of pedestrians is brief descriped. Color models and statistical models were regional template-based, people-oriented track. Using the extended kalman filter to predict a moving target may be in the area next time, narrowed the scope of the search target tracking. Make full use of the results of the moving target detection, and improve the matching efficiency and tracking speed of the target.