Arduino based Smart Fire Detection System

Abstract

The system utilizes an Arduino Mega 2560 to process signals from an infrared (IR) flame sensor providing near immediate fire detection alerts via an LCD display and a passive buzzer. To reduce false positives, the system is set to integrate both infrared and optical cameras with machine learning based object detection models. Through implementation of trained Convolutional neural networks (CNN) we are able to use the onboard camera to relay images to both servers and on-board storage for image detection and verification of any false positives that can be used for retraining.