摘要
自行式智能定位消防炮是一种适用于石油化工场所的智能消防灭火机器人,针对该类消防机器人的火灾检测,提出了一种基于等效椭圆特征和支持向量机(SVM)的红外热像仪火灾检测系统。该系统利用红外热像仪灰色图像中的最高温度作为火灾的初始判断依据,通过区域生长法对图像进行分割,提取疑似火灾区域,提取疑似火灾区域的等效椭圆特征,并利用SVM完成火灾检测过程。实验结果表明,该系统能对石油化工场所火灾进行快速、有效检测,并有良好的环境适应性。
Self-propelled intelligent positioning fire-fighting gun is a kind of intelligent fire-fighting robot designed for petrochemical industry.For the fire detection of this kind of fire-fighting robot,a fire detection system based on equivalent elliptical feature of infrared thermal imager and support vector machine(SVM)was proposed.The system uses the highest temperature feature in the gray image of infrared thermal imager as the initial judgment for the fire.Then the image is segmented by the region growing method to extract a suspected fire region.Finally,the equivalent elliptical characteristics of the suspected fire area are extracted,and the fire detection process is completed by SVM.The experimental results showed that the system can quickly and effectively detect fires in petrochemical places and has good environmental adaptability.
作者
张航
赵敏
王璐
姚敏
郭瑞鹏
ZHANG Hang;ZHAO Min;WANG Lu;YAO Min;GUO Rui-peng(College of Automation Engineering, Nanjing University ofAeronautics and Astronautics, Jiangsu Nanjing 211106, China;Tianjin Fire Research Institute of MEM, Tianjin 300381, China)
出处
《消防科学与技术》
CAS
北大核心
2018年第11期1563-1567,共5页
Fire Science and Technology
基金
国家重点研发计划项目"石油化工工程火灾应急处置特种装备研发"(2017YFC0806602)
国家自然科学基金面上项目(51875289
61873124)
航空科学基金项目(2016ZD52036)
江苏省青年基金项目(BK20150746)
