Research And Implementation Of Image Type Fire Detection And Fire Source Location System

Fire hazards is obvious for humans, prevent and control them effectively, minimize the losscaused by the fire is the focus of today’s technology research in the field of fire prevention. Thetraditional fire detectors are limited under the principle, for some large space and undergroundconstruction, common smoke, temperature sensing fire detection and alarm system is unable togather the information about the smoke temperature changes emitted by the fire quickly, while theimage fire detection technology for fire detection has the characteristics of non-contact detection,without the space height, thermal barrier, explosive, toxic and other environmental conditions,provides a new fire detection method for us.Based on the study of the basic theory of the infrared radiation, compared the existing firedetection technology, using the technology of digital image processing, designed a image type firedetection and fire source location system based on the OpenCV (open source computer visionlibrary), after the simulation and experiment, the detection system can achieve fire early warningand early protection.This paper focuses on how to detect fire and how to locate the source of fire, in order tofacilitate the initial fire extinguishing. In the fire detection technology, based on the imagecharacteristics of fire when burning in this paper, in the light of infrared temperature algorithm,axial ratio discrimination algorithm, both inside and outside flame discrimination algorithm andirregular discrimination algorithm, research on related fire recognition algorithm. Infraredtemperature algorithm is mainly based on the relationship between image pixel values with the formation temperature of the object; axial ratio discrimination algorithm is mainly based on thecharacteristics of longitudinal flame; both inside and outside flame discrimination algorithm ismainly based on the flame with internal and external stratification; irregular discriminationalgorithm is based on the irregular shape of the image characteristics of fire, and fire detectionaccuracy is further provided a guarantee reliability.In the fire source localization technology, the paper studied the fire source localizationmethod based on binocular vision. According to the binocular stereo vision ranging principle, toget through the stereo matching between parallax images, combined with the pixel coordinates ofthe flame centroid, three-dimensional reconstruction, in order to determine the three-dimensionalcoordinates of the flame, setting the foundation for large space automatic fire extinguishing.