|
|
 |
| |
|
|
|
Deep Learning-Based Forest Fire Detection Using MobileNetV2: A Comprehensive Study and Performance Evaluation |
|
|
|
PP: 131-137 |
|
|
|
Author(s) |
|
|
|
Mohamed Ben Hassen,
Kaies debbabi,
|
|
|
|
Abstract |
|
|
| Early detection and correct recognition of flames are essential to protect property and human life. This research paper seeks to study several publiclyapproaches, including deep learning-based approaches, for identifying and differentiating fires. To achieve both fire detection and classification, a unique method combining the MobileNet backbone with the Convolutional Neural Network (CNN) configuration was prposed. The obtained model performs the best according to the results, boasting an impressive 98% accuracy rate on a challenging and lucrative dataset. The use of cutting-edge technology is a significant weapon in the fight against fire catastrophe, as it immediately highlights the significance of highly effective fire detection in minimizing potential damages and/or the size of the calamity. The current work uses a noteworthy method for detecting fire occurrences in the environment under various conditions: extensive testing and analysis of the suggested CNN+MobileNetalgorithm,demonstrating excellent accuracy levels. The findings of this study provide essential building blocks for the automated fire detection system, which will raise public awareness of safety by encouraging the adoption of preventative measures in areas prone to fires. |
|
|
|
|
|
|
|
