Real-time correlogram tracking for airborne traffic surveillance

This paper presents a real-time moving object tracking system for airborne traffic video surveillance. The motion segmentation and tracking in this scenario is difficult because of the small size of the object, motion of camera and inconsistency in detected object shape. Here we present correlogram template matching system employed for tracking. Candidate objects for tracking are initially tagged by moving object segmentation. Tagged objects are then tracked using a novel integral histogram approach which provides good tradeoff between memory and processing time. Common model drift problem in the tag and track scenario is addressed by feedback mechanism.