Development of Surveillance Robots Based on Face Recognition Using High-Order Statistical Features and Evidence...

Area of Science:

• Robotics and Artificial Intelligence
• Computer Vision and Machine Learning
• Internet of Things (IoT) Applications

Background:

• Advancements in AI, CV, and IoT are transforming surveillance systems, enabling real-time processing for enhanced security.
• Mobile robots are increasingly utilized in surveillance for hazardous tasks beyond human capability.
• Existing systems face challenges with accuracy and robustness in dynamic environments.

Purpose of the Study:

• To develop a cost-effective mobile surveillance robot prototype for industrial environments.
• To integrate IoT and advanced face recognition for intelligent intruder detection.
• To enhance security by differentiating between authorized personnel and intruders.

Main Methods:

• A Raspberry Pi 4-based mobile robot equipped with a PIR sensor and camera for live data capture.
• Utilizing IoT for real-time data transmission to a control room.
• Implementing a novel face recognition algorithm combining high-order statistical features and evidence theory.
• Developing a web interface for remote robot control via Wi-Fi.

Main Results:

• The face recognition system demonstrated high accuracy (98.63%) in identifying individuals.
• The combined approach effectively addressed variations in lighting, expressions, and occlusions.
• Alert notifications with captured images were successfully sent to the control room upon detecting unfamiliar individuals.
• Experimental validation with 400 images of 40 individuals confirmed the system's effectiveness.

Conclusions:

• The developed mobile surveillance robot offers a reliable and accurate solution for industrial security.
• The integration of AI, IoT, and robust face recognition significantly improves intruder detection capabilities.
• The system's cost-effectiveness and remote control features make it suitable for diverse industrial applications.