Predictive and Probabilistic Tracking to Detect Stopped Vehicles

Many techniques and models have been proposed for vehicles surveillance in highways. In the past, tracking algorithms based on Kalman filter have been largely usedfor their efficiency in the prediction and low computationalcost. However, predictive filters can not solve long-lastingocclusions. In this paper, we propose a new mixed predictiveand probabilistic tracking that exploits the advantagesof predictive filters for moving vehicles and adopts probabilistic and appearance-based tracking for stopped vehicles. The proposed tracking is part of a complete videosurveillance system, oriented to control tunnels and highwaysfrom cluttered views, that is implemented in an embeddedDSP platform and provides background suppression,a novel shadow detection algorithm, tracking, and scenerecognition module. The experimental results are obtainedover several hours of videos acquired in pre-existing platforms of CCTV surveillance systems.