A Rule-Based Vehicle Tracking with Multiple Video Sequences
Eun-Jong Park, HyungJun So, Sunghwan Jeong, Joonwhoan Lee · 17th ITS World CongressITS JapanITS AmericaERTICO · 2007
Automatic tracking of vehicles in a video sequence is important to not only accurately estimate the speeds but also analyze the behaviors of vehicles. Because of the limited field of view, the range of visual tracking with the video sequence from a single camera is restricted. In order to enlarge the tracking range for better accuracy of speed measurements and better possibility of monitoring the vehicle behaviors, seamless tracking with multiple video sequences is necessary. This paper proposes a tracking scheme with well-synchronized multiple video sequences, in which the same rule-based tracking for each video sequence is cascaded. In each rule-base almost all possible cases that can appear in the video-based vehicle tracking are considered to construct the set of rules. Also, the rule based tracking is augmented with a Kalman filter. The resulting tracking is successful in collecting data such as temporal variations of vehicle speed and behaviors of individual vehicles in the enlarged tracking region.