Method of Identifying Fire Flames in Vision Based on the Variation Contours
Shi-lin SHEN,Li-tao CHEN,Fei-niu YUAN,Chun-yu YU,Li-feng QIAO,Yong-ming ZHANG
DOI: https://doi.org/10.3969/j.issn.1009-6094.2009.01.037
2009-01-01
Abstract:This paper is aimed at introducing a method of identifying fire flames in vision based on the variation contours.As is known,conventional point smoke and fire detectors are widely used at present fire-fighting practice to detect the presence of certain particles generated by smoke and fire by identifying their ionization and photometry.However,a big weakness of point detectors lies in that in large rooms,it may take a long time for smoke particles to reach a detector and,moreover,it is impossible to operate then in an open space.On the other hand,the vision based fire detection is a good method for large or open spaces and in turn the research in this way has become very active recently.However,the method is still facing great technical challenges since its high false alarming rate.A key problem in this way is not easy to pick up vision features.In addition,the flickering of fire flame is also a common phenomenon in favorite for detecting the fire plume and air flow.Since the contour of fire flame keeps changing with the time,it is necessary to use a CCD camera to gather the video images of fire flame.In doing so,the following methods are supposed to pre-process the image.Firstly,median-filtering system is used to remove the noise of the image.Secondly,pre-mixed background and difference models are used to detect the moving region in the video image.After that,it would be easier to pick up the contour of the image moving region.And,finally,we have brought up forward a flickering model of centroid distance of the contour of fire flame image to define the characteristics of the image,with the distance model of flickering contours between the two successive frames presented in the normalized Fourier descriptors.The results of our experiment show that the algorithm is effective for fire recognition and therefore significant for improving the accuracy of the true fire alarm and reducing the those of the false ones in the meantime time.