Fire Detection And Prevention System

Abstract

The risk of accidental fres or explosions is a matter of grave concern to humanity since it endangers human lives and threatens social safety. The system proposed in this thesis project is an advanced and low-cost solution for detecting and extinguishing fre in the early stage with minimal resources. The system is built based on the Internet of Things (IoT) platform, a modern and effective solution for connecting computing devices and machines to the Internet; and image processing using OpenCV, an open-source computer vision library. Important hardware components included in the system are NodeMCU, camera, and pump. The pump is connected with the relay and placed directly under the camera. The NodeMCU acts as a processor to control the camera, relay, and pump by receiving and processing signals transmitted through a serial connection. The system operates through two main phases: fre detection and fre extinguisher. In the frst phase, the camera repeatedly spins with a specifc angle in the horizontal direction. The system continuously processes images collected from the camera and runs pre-trained YOLO model to infer whether there is the presence of fre. In the second phase, when fre gets detected, the system program transmits signals to adjust the camera horizontally and vertically so that the water pipe points towards the fre. Then, the pump is activated, and the system begins to put out the fre.