Development of an IoT-Based Fire Recognition Using Artificial Intelligence

This paper presents the development of an IoT-based Fire Recognition using Artificial Intelligence. One of the most ubiquitous causes of worldwide deforestation and devastation of wildlife is fire. Disasters caused by fire in various residential homes, commercial and industrial locations are major concern because huge damage of infrastructure as well as human lives is always recorded. Early detection of the fire to protect valuable lives and properties is required. This research is geared towards developing an early fire recognition system capable of giving warning report in order to prevent fire incidence in the environment. The system was divided into hardware and software. The IoT hardware part of the system was carefully designed, considering all the sensors, actuators and controllers for the detection of the early fire. The computer vision was deployed to filter information from the fire images captured for early fire recognition. A sensor camera is used to monitor the environment in case of smoke or early fire, capture the image and processed using machine learning. A pre-trained Google model deployed to confirm the presence of a fire outbreak. The confirmation triggers an alarm, sends image via telegram bot to the personnel on duty and activates fire suppression process. The personnel use smart phone to view the incidence via configured telegram app. The prototype system was built and tested with good results.