Optimized Fire Classification with MobileNetV2 and CNN: Reaching 98% Accuracy

In this research paper, the optimised fir classification by the application of deep learning using the MobileNetV2 model is being traversed. Thre reliable and effective fire detection techniques undergoes development with the provided frequency and intensity of wildfires. This research examines the utilization of the MobileNetV2 model for the fire classification. The model aims to develop a robust and efficacy system that can perform in resource-restricted conditions. The proposed model was trained and appraised on an extensive dataset including images of fire and non-fire scenarios. Performance metrics determines the model high accuracy and effectiveness. Particularily, the model attained a prescios of 0.98, recall of 1.00, and an F1-score of 0.99 for fire images, while non-fire images yielded a precision of 1.00, recall of 0.90, and an F1-score of 0.95. The overall accuracy of the model is 0.98, including the macro and weighted averages of precision, recall, and F1-score all reflecting similarly high values (0.99, 0.95, 0.97 for macro avg and 0.98, 0.98, 0.98 for weighted avg, respectively). These results highlights the model's capability for the refinement of fire detection potential in real-world applications, aiding to enhanced quick response and public safety. The study coordinates with ustainable Development Goals (SDGs) by encouraging enviromental sustainability, advancing climate action, protecting life on land, enhancing disaster risk reduction, and fostering innovation in emergency response technologies.