University of Houston
Department of Computer Science


In partial fulfillment of the Requirements for the Degree of
Master of Science


Nattawat Chanjaraspong
will defend his thesis

IMPLEMENTING A LOW-COST PC-BASED SURVEILLANCE SYSTEM

WITH AUTOMATIC REAL-TIME THREAT DETECTION



Abstract



There is an increasing need to monitor and detect suspicious and potentially harmful activities in public and private facilities, even in "low-threat'' establishments such as supermarkets and university laboratories. In these places, the deployment of a surveillance system is usually constrained by a limited budget. The purpose of this thesis is to describe a cost-effective, personal-computer-based, real-time camera surveillance system that does not require human monitoring. Personal computer systems have increased their computational power, but some heavy-duty tasks such as complex image processing still cannot be performed in real-time. We propose a simple algorithm for automated real-time camera surveillance that adheres to the "cheaper, better, on-time" philosophy suitable for security-monitoring in an environment requiring a low-cost surveillance system. This system basically consists of 4 modules: (1) the moving pixel detection module uses the background subtraction approach; (2) the human body parts extraction module uses the human skin intensities and a contour-based method to extract each part and classifies them; (3) the threat assessment module uses scene knowledge to preserve the timing requirements; and (4) the timing analysis for each process to and guarantees that it will not miss the deadline.




Date: Monday, December 9, 2002
Time: 11:00 AM
Place: 550-PGH



Faculty, students, and the general public are invited.
Thesis Advisor: Dr. Albert Cheng