Kalman filtering-based modified Cam-Shift vehicle tracking algorithm for highway traffic conditions

A new Kalman filtering-based Cam-Shift vehicle tracking algorithm is proposed for video monitoring of highway traffic conditions in this paper. First, the approximate location and size of the initial search windows are determined by Kalman filtering. Then, taking the predictive values as initial values, the Cam-Shift algorithm can give information on target vehicles in real time and with a high accuracy, and improve its anti-interference ability and tracking speed. Our results show that this algorithm is capable of vehicle tracking and can detect highway traffic conditions accurately.