Skip to content
Conference Open access

Predictive Virtual Multi-Sensor Fire Detection and Adaptive Automatic Suppression System Using Edge IoT

Sep 2026 · 2026 International Conference on Secure IoT and Cybersecurity (ICOSICS) · pp. 773-780 · 0 citations · 19 references

Abstract

Industrial fires present serious threats to workers, infrastructure, equipment, and overall operations in any industrial entity, necessitating the need for reliable methods for early-fire detection and suppression. Most conventional fire detection systems rely on multiple physical sensors, which leads to increased costs, difficulties in maintenance, and deployment complexities. This thesis proposes a Predictive Multi-Sensor Fire Detection and Adaptive Automatic Suppression System Using Edge IoT to address multiple challenges. The developed approach uses a novel Virtual Multi-Sensor paradigm that leverages channels extracted from a single CCTV feed for smoke, flame, and motion detection. A novel approach of using spatial and temporal features of consecutive video frames is employed to capture the extent of events and their persistence over time and relative location for the purpose of analysis. An Adaptive Visual Evidence Fusion technique fuses diverse evidence types, including motion, smoke, and flame, into a single metric that represents the confidence level about the possibility of fire. After determining the current state of the fire, the system implements a Fire Progression Prediction algorithm to predict the most likely short-term progression state of the fire. The current and likely state of fire are then used to trigger adaptive responses that range from monitoring, warning, and alarm activation to suppressing the detected fire. An Edge-IoT interface based on ESP32 is used for low-cost local actuation and monitoring. The proposed approach can be used in manufacturing units, warehousing, chemical, power plants, laboratories, and other industrial facilities that have CCTV surveillance systems installed and are interested in leveraging existing infrastructure to enable automatic monitoring of fires and adaptive response. The proposed work contributes to United Nations Sustainable Development Goals (UNSDG) 8 and 9.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.