Smart Fire Surveillance and Alarm Generation System Using Raspberry Pi and Yolo-Based Detection

The proposed Smart Fire Surveillance and Alarm Generation System successfully demonstrates a reliable, low-cost, and intelligent approach to real-time fire detection by integrating YOLOv8-based computer vision, OpenCV image processing, and ESP8266-based sensor communication on a Raspberry Pi 4 platform. The hybrid detection mechanism, combining visual analysis from a USB webcam with physical verification from MQ-2 smoke and IR flame sensors over Wi-Fi, effectively addresses the core limitations of conventional fire alarm systems, including delayed detection, high false alarm rates, and the absence of remote monitoring capabilities. Experimental results confirm that the system achieves a detection accuracy of 89.4%, an average response time of 2.3 seconds, and a false alarm rate of 4.7%, representing significant  improvements  over  traditional  sensor-only approaches. The three-tier decision logic ensures proportional and reliable alert generation, while the Flask and React.js-based web dashboard enables live surveillance and remote accessibility over a    local  network.             The     distributed architecture, with the ESP8266 handling sensor acquisition independently of the Raspberry Pi's main processing pipeline, further improves system responsiveness and scalability. Future work will focus on integrating thermal imaging to improve detection accuracy under dense smoke conditions, deploying ONNX Runtime for optimised model inference on the Raspberry Pi, and extending the system to support multi- node fire zone monitoring with a centralised alert management interface.  

[1]  W. Zheng, Z. Li, Y. Liu, and C. Li, "A review of fire detection methods based on deep learning," IEEE Access, vol. 10, pp. 64760–64781, 2022. (Supports: traditional fire detection limitations, photoelectric /ionisation detectors) [2]  K. Muhammad, J. Ahmad, I. Mehmood, S. Rho, and S. W. Baik, "Convolutional neural networks based fire detection in surveillance videos," IEEE Access, vol. 6, pp. 18174–18183, 2018. (Supports: false alarms, environmental triggers, emergency response costs) [3]  B. U. Toreyin, Y. Dedeoglu, U. Gudukbay, and A. E. Cetin, "Computer vision based method for real-time fire and flame detection," Pattern Recognition Letters, vol. 27, no. 1, pp. 49–58, 2006. [4]  K. Muhammad, J. Ahmad, and S. W. Baik, "Early fire detection using convolutional neural networks during surveillance for effective disaster management," Neurocomputing, vol. 288, pp. 30–42, 2018. [5]  K. Muhammad, M. Sajjad, F. Mehmood, and S. W. Baik, "Efficient fire detection for uncertain surveillance environment," IEEE Transactions on Industrial Informatics, vol. 15, no. 5, pp. 3113–3122, 2019. . [6] T. Celik and H. Demirel, "Fire detection in video sequences using a generic colour model," Fire Safety Journal, vol. 44, no. 2, pp. 147–158, 2009. [7]  G. Bradski and A. Kaehler, Learning OpenCV: Computer Vision with the OpenCV Library, 1st ed. Sebastopol, CA: O'Reilly Media, 2008. [8]  K. Muhammad, J. Ahmad, Z. Lv, P. Bellavista, P. Yang, and S. W. Baik, "Efficient deep CNN-based fire detection and localization in video surveillance applications," IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 49, no. 7, pp. 1419–1434, 2019. [9]  Y. LeCun, Y. Bengio, and G. Hinton, "Deep learning," Nature, vol. 521, pp. 436–444, May 2015. [10]  J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, "You Only Look Once: Unified, Real-Time Object Detection," in Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, 2016, pp. 779–788.