简介:Visualfiredetectiontechnologiescandetectfireandalarmwarningsearlierthanconventionalfiredetectors.Thisstudyproposesaneffectivevisualfiredetectionmethodthatcombinesthestatisticalfirecolormodelandsequentialpatternminingtechnologytodetectfireinanimage.Furthermore,theproposedmethodalsosupportsreal-timefiredetectionbyintegratingadaptivebackgroundsubtractiontechnologies.Experimentalresultsshowthattheproposedmethodcaneffectivelydetectfireintestimagesandvideos.ThedetectionaccuracyoftheproposedhybridmethodisbetterthanthatofCelik’smethod.