Autonomous Blazebot: A Real-Time Fire Detection and SMS Alert System Using AI and GSM Technology

Indoor fire incidents pose a significant threat to both life and property, particularly in areas that are not regularly monitored or are isolated. Traditional fire detection systems, which typically rely on smoke, temperature, or gas sensors, tend to be passive, prone to false alarms, and incapable of providing intelligent, real-time alerts to users in remote locations. This paper addresses these challenges by introducing the development of an autonomous Blazebot for fire detection and alert systems using GSM technology. The proposed system integrates a lightweight YOLOv8n (You Only Look Once version 8 nano) deep learning model, deployed on a Raspberry Pi 4, to continuously recognize flames visually through a USB camera. Once a flame is confirmed, the system sends an alert signal to an Arduino Mega 2560, which then activates a SIM900 GSM module to send SMS notifications to designated recipients. The system successfully identified flame sources with a minimum size of 2500 pixels at distances up to 350 cm, achieving optimal accuracy between 200 and 250 cm. The average delay for SMS transmission was recorded at 10.01 seconds after detection. These findings demonstrate the viability of a costeffective, real-time, vision-based fire detection and communication system suitable for settings without internet access or constant human oversight.