A Low-Power Low-cost System for Disaster Locations Detection using ESP32 CAM and TinyML

This paper presents a low-cost, low-power solution for real-time natural disaster or calamity location detection using the ESP32 CAM, a GPS module, and TinyML. By overlaying calamity locations on open-source street maps, our system improves user interaction with spatial data, aiding navigation and safety during disasters. The model achieves 88% accuracy while using minimal RAM and FLASH memory, offering a cost-effective alternative to expensive devices. Additionally, it can be deployed on UAVs, or drones, or towers, for extensive scanning, alerting remote servers or disaster rescue centers upon detection for timely responses.