计算机工程与应用Issue(13)：156-159,4.DOI:10.3778/j.issn.1002-8331.1111-0038
基于神经网络与加权融合的火灾火焰识别研究
Research of fire detection algorithm based on neural network and weighted fusion
摘要
Abstract
For the indoor environment as the application background,this paper designs the feature fusion flame detection algo-rithm with static and dynamic flame features,which gives the quick decision to the fire flame.Firstly,the paper makes the mo-tion detection to the video image,extracts the color feature,and creates a new color criterion in HIS space,then gets round and peak number of fire flame area.Finally,it researches on the characteristics of the flame flicker,taking all these features as neural network input,and ultimately uses of weighted fusion algorithm to determine whether it is the fire flame.关键词
运动检测/火焰识别/阈值分割/特征融合Key words
motion detection/flame detection/threshold estimation/feature fusion分类
信息技术与安全科学引用本文复制引用
靳琪琳,段锁林..基于神经网络与加权融合的火灾火焰识别研究[J].计算机工程与应用,2013,(13):156-159,4.基金项目
常州市科技计划项目（No.CJ20110023）。 （）
