Fire suppression system based on MATLAB video processing and FPGA

Abstract

Class A fires (fires out of wood, paper, cloth…etc) are the most common firebreaks in confined space, such as houses, desks, schools…etc. In this project a quick and reliable automatic fire suppression system is designed and implemented. It uses a camera to monitor the confined space at all times. Image processing techniques are used to treat the video stream and starts an alarm whenever a fire is detected. The system uses RGB components to detect the RGB characteristics of fire, and uses luminance and color information stored in YCbCr components to confirm the rules of detection. The system can locate the fire with high precision based on the geographic location of camera and location of fire pixels in the image. After locating the fire an FPGA board is instructed to point a water cannon towards the fire source. But first Confirms the fire using a flame sensor, then an alarm is started and water is ejected from the cannon. The whole process takes matter of some seconds before action is taken. The system has proved itself to be very accurate in small monitored spaces as traced in the objectives. Further improvement have to be made before it can be installed in almost any institution where class A fires can be a risk.