Research Of Control And Fire Identification System For Inteligent Fire

Fire is one of the most serious and frequent disasters, and it will do harm to people’s life and property.The fire alarm and firefighting have becoming more and more important with the continuous developmentof the society and economy. According to experience from studying the fire,, people have also producedmany effective fire products by means of scientific research and development. Especially, the fire flamedetection system based on video image is proposed in recent years. It has attracted much more attention ofnumerous researchers at home and abroad and been applied to some complicated environment because ofits rapid and accurate detection performance.This paper designs a video image fire flame detection system based on BP artificial neural network.Firstly, pre-processes the video image effectively based on image processing, highlights the usefulinformation of video image and separates the suspected area of fire flame image accurately. Secondly,adopts the video image processing and analysis method, extracts five feature information about thecharacteristics of the fire flame including the expansion area, flicker frequency, circular degreecharacteristics, angle characteristics and bar centric coordinates rate according to the static and dynamiccharacteristics of the fire flame. Finally, a three layers BP artificial neural network structure has been set upto realize recognition and judgment for the fire by taking into account the five characteristic information offire flame.The experiment results indicate that the video image fire flame detection system based on BP artificialneural network proposed in this paper has higher anti-jamming ability and can identify the fire flame andfire alarm accurately.