Paper Contents
Abstract
Fire accidents are a widespread and emerging problem in the majority of industries, such as domestic, industrial, and environmental settings. The shortcomings of conventional fire detection deviceslike diminished rates of response, dependency on fixed thresholds, and inability to cope with dynamic environmental inputsrequire smart and forecast-based detection technologies. This paper introduces an IoT sensor-based deep learning-penetrated fire detection system aimed at identifying potential fire hazards in their initial stages with high reliability and few false alarms.The system architecture combines several IoT sensors installed in target environments to obtain real-time data regarding temperature, humidity, concentration of smoke, carbon monoxide (CO), carbon dioxide (CO), and intensity of light. The multi-modal data streams are sent to a cloud platform or central processing system, where they are analyzed via state-of-the-art deep learning models like Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks. LSTM models are used for their capacity to encode temporal relationships in sequential sensor data, while CNNs aid spatial pattern detection and anomaly discovery.
Copyright
Copyright © 2025 Vandana Reddy N. This is an open access article distributed under the Creative Commons Attribution License.
