Building Fire Positioning Algorithm Research Based On Image Processing

Fire image automatic monitoring technology is based on the advantages of imageprocessing and computer language, high-speed computing, and proposes the use of a flameimage information to analyze and judge the fire monitoring techniques. And implementedbased on the image of the fire alarm system, this paper, the system does three studies.First, due to fire flame recognition algorithm is complex and demanding environment isproposed based on multi-feature flame image recognition method. Through the imagedifference method to detect moving flame suspected area, then the detected region is set to aregion of interest, and then through the region of interest within the target suspected flametexture and shape features of image analysis, exclusion does not have a flame color objects,further judgment decisions, and ultimately determine whether a flame produced.Experimental results show that this method has a higher detection rate, can reduce the falsealarm rate and the anti-interference ability.The second part of the research article through binocular calibrated three-dimensionalpositioning the flame position, followed by several of the camera calibration methods werestudied, and finally the image coordinates of the center of the fire and the geometriccharacteristics of the calculation method parameter calculation, and proposed a An improvedfire image center coordinates calculated.Finally, the design of an image based on image processing type fire alarm system. Systemhardware design, including the overall composition of the system and hardware components;system software design, including software design requirements, design of tender,requirements analysis, software development language of choice and the overall structure;Finally this paper, system design, conducted a fire detection simulation.