Intelligent Fire Surveillance: A Deep Learning Framework for Accurate Indoor Fire Detection

Fire is a naturally occurring but destructive phenomenon capable of causing extensive property damage and loss of life. Early detection is crucial for mitigating fire hazards, as it helps prevent the spread of fire, saves lives, and min-imizes infrastructure damage. Traditional fire detection often face challenges including low sensitivity, difficulties in large-scale deployment, high false alarm rates, and detection delays. This research presents an innovative fire detection system that integrates a Deep Learning model framework to monitor, detect, and warn about fires in indoor environments like buildings, rooms, and restaurants.The framework utilizes a pre-existing Convolutional Neural Network (CNN), specifically leveraging the VGG16 model, as the basis for its Deep Learning architecture. Through transfer learning, we develop a FireNET architecture specifically designed for fire detection tasks with an accuracy of 99.50%. This integration of Deep Learning in fire detection provides real-time detection with high speed and accuracy and effectively overcoming the limitations of traditional fire detection sensors.