ABSTRACT

ABSTRACT: The system that can perform the human detection and population statistics real-timely according to the video scene of the fixed color camera in intelligent video monitoring system is realized. The system’s interface is designed by the MFC, and we extract moving foreground to detect a human movement through the background difference method. The HOG is used to employ the human head and shoulders’ feature extraction. The linear support vector machine as classifier realizes the head and shoulder model classifying. Dynamic human body counting is finally realized through targets mark and the real-time tracking. The experimental results show that the system can accurately realize the quantity statistics of floating population.