Answers are grounded in this paper and cite their sources.
Early fire detection and rapid suppression are critical to minimizing risks in residential and industrial settings. This study presents a compact smart fire safety system that integrates real-time deep learning-based detection with automated extinguishing. The lightweight YOLO11n model, trained over 300 epochs, achieved a mean average precision of 0.94 and successfully detected small-scale fire outbreaks occupying only 5–15 % of the video frame. A Windows-based interface developed in Flet connects to an 8 MP IP camera for continuous monitoring, while a Raspberry Pi 4-controlled servo–stepper motor system guides real-time nozzle positioning. Low-latency communication was established via Google Firebase (200 ms write, 100 ms read), ensuring synchronized operation. Experimental trials comparing water spray and jet modes showed that spray was more effective, achieving a maximum extinguishing ratio of 0.5 at 0° nozzle angle versus 0.3 for jet. The system successfully extinguished fires up to 5 m distance and 1.8 m height, with efficiency declining at wider nozzle angles (±25° and ±50°). The results demonstrate the novelty and practicality of an AI-driven detection and suppression framework, offering a scalable and modular solution for smart buildings, industrial plants, and other safety-critical environments. • YOLO11n model achieved 0.94 mAP for real-time early fire detection. • Automated nozzle actuation via Raspberry Pi ensured precise suppression. • Water spray mode outperformed jet, achieving ER of 0.5 at 0° angle. • System successfully extinguished fires up to 5 m distance and 1.8 m height. • Compact, modular AI design enables scalable deployment in smart buildings.
No takes yet. Share an insight, caveat, or question.
Yuvaraj et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: