Fire Detection and Segmentation using YOLOv5 and U-NET
Wided Souidene Mseddi, Rafik Ghali, Marwa Jmal, Rabah Attia
Abstract
The environmental crisis the world faces nowadays is a real challenge to Human Beings. One notable hazard for humans and nature is the increasing number of forest fires. Thanks to the fast development of sensors and technologies as well as computer vision algorithms, new approaches for fire detection are proposed. However, these approaches face several limitations that need to be resolved, precisely, the presence of fire-like objects, high false alarm rate, detection of small size fire objects, and high inference time. An important step for vision-based fire analysis is the segmentation of fire pixels. Hence, we propose, in this paper, a novel architecture, combining YOLOv5 and U-net architectures, for fire detection and segmentation. Using a dataset of wildland fires mixed with fire-like object images, the experimental results proved that the novel architecture is reliable for forest fire detection without false alarms.