基于烟雾动态特征分析的实时火灾检测  

Smoke Dynamic Features Based Real-Time Fire Detection

作　　者：王欣刚[1] 魏峥[1] 刘东昌[1] 郑晓隆[1] 

机构地区：[1]中国科学院自动化研究所智能感知与识别实验室,北京100080

出　　处：《计算机技术与发展》2008年第11期9-12,17,共5页Computer Technology and Development

基　　金：国家科技支撑计划(2006BAK31B03)

摘　　要：传统的离子式、光电式、吸烟式等烟雾探测器,需要非常靠近烟雾源时才能完成检测,不能应用于室外大空间、远距离等特殊场合的火灾监控系统中。提出了一种基于烟雾动态特征的图像型火灾识别算法。通过监测普通CCD摄像机拍摄的视频图像,进行运动物体的检测分割,分析图像帧中运动物体的动态特征,进而判断监控区域是否有火灾烟雾的发生。实验结果表明,该算法具有比较好的鲁棒性,能够有效地降低火灾监控系统的误报漏报率。The traditional ionization smoke detector, photoelectric smoke detector and aspirated smoke detector etc were unable to be used on the fire detection to the large and open spaces. A fire detection method based on smoke dynamic features is proposed. After the detection of the moving objects through the video sequence of the normal CCD eameral, the method analyzed the dynamic features of the moving objects, which determined the region whether have the fire smoke occurrence or not. Experimental results show that the method has the characteristic of robusticity and the ability to reduce the false and lose alarms of the fire surveillance system.

关 键 词：动态特征 分割 火灾探测 烟雾图像 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]