Abstract:
The delineation of a fire detection system for its use in indoor environments is scrutinized in this thesis. The Internet of Things (IoT) is an advancing field of technology where a group of physical devices or devices with software, processing capability, sensors, and other technologies exchange data and connect with other devices over the web or other communication networks. Fire Rooster is an IoT-based fire detection system using machine learning (ML). The main agenda of this system is to distinguish between hazardous fire and non-hazardous fire and thus inform the remote user in case of a fire accident. The potential areas for deployment of this system are offices, houses, stores, libraries, labs, and malls. This thesis examines the solutions and shortcomings of the already present system. The updated system distributes and includes features like object detection, record maintenance, and message conveyance to the owner. The camera will take the live feed, and the ML model deployed on the PyCharm will distinguish between hazardous and non-hazardous fire. If a hazardous fire is detected, the message will be delivered to the owner via SMS on sim. Moreover, the alarm will also be generated using buzzer at the scene of the accident.
