Thermal Infrared Sensing for Near Real-Time Data-Driven Fire Detection and Monitoring Systems.

Area of Science:

• Robotics
• Computer Vision
• Fire Safety Engineering

Background:

• Mobile robotics are increasingly used for hazardous environment monitoring, including fire detection.
• Effective fire detection requires robust sensing and image processing, especially in extreme conditions.
• Autonomous systems for early fire warning and monitoring are crucial for public safety.

Purpose of the Study:

• To evaluate the sensing capabilities and image processing pipeline of thermal imaging sensors for fire detection.
• To develop data-driven thermal imaging analysis methods for robotic perception in fire scenarios.
• To provide insights for integrating thermal cameras into autonomous unmanned aerial vehicle (UAV)-based fire surveillance systems.

Main Methods:

• Overview of image processing algorithms for thermal imaging (data compression, enhancement).
• Development of a data-driven analysis method for fire situation awareness in robotic perception.
• Experimental study of thermal camera behavior in controlled fire scenarios with in-depth data analysis.

Main Results:

• Identified key image processing techniques for thermal imaging in fire detection.
• Demonstrated a data-driven approach for analyzing thermal data to improve fire situation awareness.
• Revealed the operational characteristics of thermal cameras in fire environments through experimental testing.

Conclusions:

• Thermal imaging is a viable sensor technology for robotic fire detection systems.
• Data-driven analysis and specific image processing techniques enhance thermal camera utility in robotic perception.
• Key considerations for integrating thermal cameras into UAV-based fire surveillance were identified, enabling autonomous early warning systems.