HUMAN DETECTION AND TRACKING ENHANCING SECURITY SYSTEMS AT PORTS OF ENTRY

Abstract

The dissertation undertakes the critical application of establishing smarter surveillance systems to improve security measures in various environments. Human detection and tracking are two image processing methods that can contribute to the development of a smart surveillance system. These techniques are used to identify and detect moving humans in a surveyed area. The research enables the incorporation of personnel detection and tracking algorithms to enhance standard security measures that can be utilized at ports of entry where security is a major hurdle. This system allows authorized operators on any supported console to monitor and receive different alerts levels to indicate human presence.

The presented research focuses on two human detectors based on the histogram of oriented gradients detection approach and the Haar-like feature detection approach. According to the conducted experimental results, merging the two detectors, results in a human detector with a high detection rate and lower false positive rate. A novel approach to use both detectors is proposed. This approach is based on a feedback messaging system that inputs parameters from both detectors to output better detection decisions. An object tracker complements the detection step by providing real-time object tracking. An alert system is also proposed to automatically report potential threats occurring in the surveyed area.