Fire Detection Model Based on the Yolov8 Model for the Industrial Field

Document Type : Original Article

Authors

1 Department of Artificial Intelligence, Benha University, Cairo, Egypt.

2 Scientific Computing Department, Benha University, Cairo, Egypt

3 Scientific Computing, benha university(Faculty of computers and artificial intelligence), zifta, egypt

4 Benha University

Abstract

This paper proposed a model to detect fires in the industrial field using You Only Look Once Version 8 (YOLOv8) framework. The proposed model is based on three primary stages which are data pre-processing, feature selection, and evaluating the results using a variety of metrics. Images are resized, enhanced, noise is reduced, and videos and images are labeled with bounding boxes surrounding the fires during the data pre-processing step. The YOLOv8 model's speed and accuracy make it the preferred choice for feature selection. The performance of the proposed model is assessed using a variety of metrics, including accuracy, precision, and recall. Furthermore, the suggested model is trained in a real-time system that is capable of processing camera feeds in real time. When a fire is detected, the building's fire alarm should go on to alert people and tell them to evacuate. The experiment's findings show that the recommended model produced results with 98.1% accuracy, 98.9% precision, 95.3% recall, and 98.1% mAP. Finally, the proposed model is contrasted with existing methods on the same dataset.



يتم تغيير حجم الصور وتحسينها ، ويتم تقليل الضوضاء ، ويتم تصنيف مقاطع الفيديو والصور مع صناديق محيطة تحيط بالحرائق أثناء خطوة المعالجة المسبقة للبيانات.
 

Keywords

Main Subjects

Files

Share

How to cite

Statistics