YOLO Based Real time Forest Fire Detection

Authors

  • Shireesh Kumar P A, Keerthi Reddy, Surekha Thota, Shreya Merin, Thejas HR

Abstract

Forest fires are very hazardous and cannot be ignored knowing the fact that they have the potential to tear down the balance between flora and fauna in the distant future. In this paper the discussion related to a popular model based on AI i.e. YOLO (You Only Look Once), used for the detection of the fire outbreaks those which mainly involve Wildfires is put together. To generalize, the power of AI is incorporated to detect and help in taking early precautions from the immense damage that could be caused by the wildfire outbreaks. The paper begins with introduction initially comprising with history of fire outbreak incidents, causes and its after effects. The literature survey references lead to deep understanding of previous models and its limitations which are key points for further enhancements. Our outline work has been explained in phases: dataset pre-processing, environment setup and YOLOv3 model training and validation. By exploiting YOLOv3 this paper proposes an algorithm which is capable of detecting real time scenario which in our case is forest fire with required precision.

Keywords: YOLOv3, Real-time detection, Deep Learning, Wildfire

Downloads

Published

2020-05-16

Issue

Section

Articles