In this paper, a system for real-time object recognition and tracking for r
emote video surveillance is presented. In order to meet real-time requireme
nts, a unique feature, i.e., the statistical morphological skeleton, which
achieves low computational complexity, accuracy of localization, and noise
robustness has been considered for both object recognition and tracking. Re
cognition is obtained by comparing an analytical approximation of the skele
ton function extracted from the analyzed image with that obtained from mode
l objects stored into a database. Tracking is performed by applying an exte
nded Kalman filter to a set of observable quantities derived from the detec
ted skeleton and other geometric characteristics of the moving object, Seve
ral experiments are shown to illustrate the validity of the proposed method
and to demonstrate its usefulness in video-based applications.