Author(s):
1. Hoor Ul Ain Tahir:
Bahria University Islamabad,Pakistan
2. Abdullah Waqar:
Bahria University Islamabad,Pakistan
3. Shehzad Khalid:
Bahria University Islamabad,Pakistan
4. Syed Muhammad Usman:
Bahria University Islamabad,Pakistan
Abstract:
Wildfires are one of the most expensive and lethal natural disasters on the planet, destroying millions of hectares of forest resources and endangering the lives of people and animals. Such accidents are time-sensitive and can result in significant loss of life and property if not dealt with timely. Detection of fire at an early stage using aerial videos can reduce personal and property losses. This research focuses on the detection of fire locations monitored by UAV drones. Predicting fire behavior can assist firefighters in improved fire management and forecasting for future events, as well as lowering the firefighters' risk to life. Recent advancements in aerial imagery suggest that these images are valuable in the detection of wildfire. Drones and Unmanned Aerial Vehicles (UAVs) are among the different methods and technology for aerial imagery that is being used to obtain information about the fire. We present a YOLOv5 based deep learning model for fire detection. The proposed method detects fire in a real-time environment with high accuracy by evaluating a video frame by-frame to detect such anomalies in real-time and sends a warning to the relevant authorities. In terms of detection performance, our technique outperforms existing fire detection systems. On the FireNet and FLAME aerial picture datasets, we evaluated the proposed method's performance and achieved the F1-score of 94.44%.
Page(s):
1-1
DOI:
DOI not available
Published:
Journal: IEEE International Conference on Digital Futures and Transformative Technologies (ICoDT2) May 24-26, 2022 (Book of Abstracts), Volume: 1, Issue: 1, Year: 2022
Keywords:
deep learning
,
Wildfire Detection
,
Aerial Images
References:
References are not available for this document.
Citations
Citations are not available for this document.