基于火焰视觉虚化性的多特征融合火焰检测  

MULTI-FEATURE FUSION FLAME DETECTION BASED ON VISUAL BLURRING OF FLAME

作　　者：徐兵荣 叶炯耀 Xu Bingrong;Ye Jiongyao(School of Information Science and Engineering,East China University of Science and Technology,Shanghai 200237,China)

机构地区：[1]华东理工大学信息科学与工程学院,上海200237

出　　处：《计算机应用与软件》2022年第11期160-167,共8页Computer Applications and Software

基　　金：国家自然科学基金项目(61872143)。

摘　　要：火焰是火灼热发光的气化部分,该部分能被可见光捕获但却会被激光穿透,将这一个特征定义为火焰的视觉虚化性。通过增加对火焰视觉虚化性的检测,提出一种新型的多特征融合火焰检测算法。该算法利用RGB-HIS颜色模型分割疑似火焰区域,同时通过双目摄像头与激光测距捕捉火焰的视觉虚化性特征,并结合了火焰的颜色、相似度、质心运动特征。采用MES(Multi-expert system)多特征融合方法构建一种高精度的火焰检测系统。实验结果表明,该算法误判率降低了9百分点,漏检率降低至0.46%,是视觉火焰检测技术很好的补足,解决了低成本商用视频火焰检测技术的难题。Flame is the glowing gasified part of fire,which can be captured by visible light but penetrated by laser.This feature is defined as the visual blurring of flame.By increasing the detection of the visual blurring of flame,this paper proposes a novel multi-feature fusion flame detection algorithm.The algorithm used the RGB-HIS color model to segment the suspected flame region.And it simultaneously captured the visual blurring features of the flame through binocular camera and laser ranging,and combined the color,similarity and centroid motion features of the flame.A high precision flame detection system was realized by using the MES multi-feature fusing method.The experimental results show that the misjudgment rate of this algorithm is reduced by 9%,and the missed detection rate is reduced to 0.46%.It is complement to the visual flame detection technology and solves the problem of low-cost commercial video flame detection technology.

关 键 词：火焰检测 火焰视觉虚化性 颜色模型 相似度 质心运动 多特征融合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]