ELEVIEW : an active elevator monitoring vision system

• Main Author: Xiao, Ping
• Other Authors: Maylor Karhang Leung
• Format: Theses and Dissertations
• Language: English
• Published: 2010
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Computer hardware, software and systems
• Online Access: http://hdl.handle.net/10356/42554
• Institution: Nanyang Technological University

In this report, a new application for active vision system in elevator monitoring, ELEVIEW, has been investigated. The research is stimulated by the reported crimes that happen inside elevators. The main goal is to investigate techniques for monitoring scene and understand actions that are occurring in an elevator. It should filter out normal actions and signal an alarm once abnormal events are detected. ELEVIEW could reduce the burden of supervision and rates of false alarms. Senior citizens, women and children can benefit immensely by this research. This study concentrates on the analysis of the scene analyzer module of ELEVIEW. The scene analyzer consists of three parts: segmentation, feature extraction and scenario analysis. The "difference picture" technique is utilized to generate two different types of pictures in segmentation. Spatialtemporal features, such as occupants' location and area, movement aggressiveness and head tops, are extracted from the images. The scenario representation is proposed to describe the activities within an elevator and perform activity classification analysis into four kinds of scenario scripts. The experiments show that the methods and algorithms employed are sufficient to classify normal scenario and certain suspicious scenarios. The results have demonstrated the potential of the prototype system to monitor an elevator automatically.