Firefighting: Applying Deep Learning Algorithms for Real-time Fire Detection

Wildfires in North America present a grave danger, leading to extensive damage. In this paper, we proposed a deep learning method to rapidly detect wildfires using drones or ground cameras in remote locations. The performance of the proposed system was assessed on public datasets, focusing on accuracy and energy efficiency. The study explores the possibility of running the proposed algorithm on low-power devices like NVIDIA Jetson Nano for real-time fire identification in remote regions.