Improved Real-Time Fire Warning System Based on Advanced Technologies for Visually Impaired People

Area of Science:

• Computer Science
• Artificial Intelligence
• Fire Safety Engineering

Background:

• Early fire detection is critical for human safety and environmental protection, especially in indoor environments.
• Existing methods often lack the speed and accuracy required for timely intervention, posing risks to property and lives.
• Assisting blind and visually impaired (BVI) individuals during fire emergencies necessitates specialized detection and notification systems.

Purpose of the Study:

• To develop an artificial intelligence-based, vision-enabled system for early flame recognition and notification.
• To enhance fire safety for BVI individuals by providing accurate, real-time information about indoor fire incidents.
• To automate manual processes, improving the efficiency and quality of fire classification.

Main Methods:

• Implementation of a vision-based early flame recognition and notification approach.
• Utilizing the YOLOv5m model for real-time monitoring and enhanced detection accuracy of indoor fire disasters.
• Automating all previously manual processes within the fire alarm control system.

Main Results:

• The proposed system demonstrated high speed and accuracy in detecting and notifying catastrophic fires.
• Successful real-time monitoring and detection capabilities were achieved, irrespective of fire shape, size, or time of day.
• Performance evaluation matrices confirmed the system's competitiveness against conventional fire-detection methods.

Conclusions:

• The developed AI-powered system effectively addresses the need for rapid and accurate early fire detection.
• The system significantly improves safety for BVI individuals by providing timely and precise fire scene information.
• The automated, vision-based approach offers a robust solution for indoor fire disaster management.