Enhancing Fire Safety: Encoder-Decoder Based Detection and IoT-Driven Response System

Innovative technologies are crucial for improving fire detection and response systems. An embedded system has been developed to improve fire detection and response by integrating a user mobile application, an admin website, and an IoT-based robot. The mobile application uses the Encoder-Decoder architecture to identify fires from user-captured images, enabling real-time transmission of critical data via the Firebase Realtime Database. The encoder-decoder model performs better with an accuracy of 97.50 %, precision of 97.20 %, recall of 97.10%, and F1-score of 97.05%. The admin website serves as a centralized management platform, utilizing Google Maps API for real-time robot tracking and video streaming from the robot's camera. The system also employs fire and person detection(YOLOv4) algorithms for enhanced monitoring. The robot's construction includes fireproof materials for resilience in fire-prone environments. This comprehensive system offers a robust solution for effective fire detection and response, improving safety and efficiency in emergencies.