On-road vehicle detection based on color segmentation and tracking using Harris-SIFT

This paper proposes a novel method for vehicle detection and tracking using a vehicle-mounted monocular camera in an intelligent vehicle system. Speed-based Adaptive Perception Zone (APZ) is first defined to ensure that the vehicle minimizes the spatial extent of the region it perceives according to its own speed. Vehicle candidates are generated using brake lights detection through color segmentation method and verified by a rule-based clustering approach. A tracking-by-detection scheme based on Harris-SIFT feature matching is then used to learn the template of the detected vehicle on line, localize and track the corresponding vehicle in live video. Our system was validated in real conditions in our prototype vehicle with state-of-the-art performance, equivalent and sometimes surpassing other methods recently published.