基于支持向量机的火灾探测系统研究

火灾的早期探测技术长期得到火灾预防专家的关注，基于图像型的火灾探测系统成为研究的重点．针对目前的火灾早期探测技术存在的问题，提出基于火灾火焰图像和支持向量机识别原理的火灾识别技术，系统获取火灾初期火焰的六个主要特征，采用基于线性核函数等不同核函数的支持向量机技术进行识别实验，得出基于线性核函数识别火灾的准确率最高的结论．本探测系统一定程度上弥补传统监测系统的缺陷，缩短预报时间，降低误报率．

Study on a Fire Image Detection System Based on Support Vector Machine

A fire early detection technique is significant to the fire prevention experts,most of research on this field is focused on the fire detection system based on image.Aming at many problems in the fire early detection techniques,a fire detection technique based on fire flame image and SVM principle is proposed,the six characteristics of fire early flame is got,using different kernel function of SVM,and the best detection method is the SVM based on linear core function.The system can remedy the defects of the ...