A System for the Detection of Persons in Intelligent Buildings Using Camera Systems-A Comparative Study

Area of Science:

• Computer Vision
• Embedded Systems Engineering
• Artificial Intelligence

Background:

• Intelligent building technologies require efficient person detection systems.
• Existing systems may lack cost-effectiveness, power efficiency, or simplicity.
• Need for adaptable person detection for various indoor environments.

Purpose of the Study:

• To design and implement a prototype of an efficient Low-Cost, Low-Power, Low Complexity (L-CPC) image recognition system for person detection.
• To evaluate static and dynamic background processing methods for person detection.
• To optimize the system configuration using the Saaty method for intelligent building applications.

Main Methods:

• Development of an L-CPC image recognition system for person detection.
• Implementation of static and dynamic background subtraction techniques.
• Comparison with the Horn-Schunck algorithm using optical flow principles.
• Evaluation of detection configurations using the Saaty method.
• Testing on simulated indoor human activities across 12 video sections.

Main Results:

• The L-CPC system demonstrates efficient person detection capabilities.
• Static and dynamic background methods show varying effectiveness depending on the scenario.
• The Saaty method successfully identified optimal system configurations.
• The system proved effective in simulated intelligent building environments.

Conclusions:

• The developed L-CPC system is a viable solution for person detection in intelligent buildings.
• The choice between static and dynamic background methods impacts performance.
• The Saaty method is effective for optimizing embedded vision system configurations.
• The prototype meets the requirements for intelligent building applications.