An adaptive Kalman predictor applied to tracking vehicles in the traffic monitoring system

Video object tracking is an important method of traffic data collection in ITS. This paper implements an approach for detecting traffic objects in urban traffic scenes by means of feature-based reasoning on visual data, and tries to track and classify traffic objects with self-defined features. An adaptive Kalman prediction algorithm is presented to improve the prediction accuracy of location. Experimental results are also shown to demonstrate the effectiveness of the proposed algorithm.