A Novel Vehicle Tracking Method on Infrared Image Processing and Traffic Jam Detection
Shin’ichi Miyamoto · 19th ITS World CongressERTICO - ITS EuropeEuropean CommissionITS AmericaITS Asia-Pacific · 2012
In the case of vehicle occlusion occurs in a camera image such as under traffic jam, it has been difficult for image processing to track each vehicle individually and accurately. Conventional method like Kalman filter is getting worse if the lack of vehicle position continues for a long time. To overcome these difficulties, the authors propose a novel tracking method which has two tracking mode and transits each other. Furthermore the authors introduce a kind of entity for tracking which is controlled by a tracking element. The authors will show experimental results by applying our method to traffic jam detection and conclusion.