摘要
对于目前火灾探测方法中存在检测率低、误报率高的普遍现象,提出了一种基于融合特征与支持向量机(SVM)的视频火焰检测算法。使用一种改进自适应混合高斯建模法获得视频里的运动目标,并结合火焰颜色模型分离出火焰疑似区域;获取疑似区域的动态、几何、纹理等特征;融合特征量,利用已训练的SVM完成识别。在测试视频集上的实验结果表明,该算法检测效果好,且耗时短。
Aiming at the problem of low detection rate and high false-alarm rate in current fire detecting method,an algorithm for video flame detection based on fusion feature and support vector machine(SVM)is proposed.Moving target is obtained by an improved adaptive Gaussian mixture modeling method,the candidate flame regions are segmented combined wiTHflame color model;the texture,geometric and dynamic characteristics of the suspected flame regions are obtained.The above characteristic quantity is fused,and fire recognition is achieved by the trained SVM before.Experimental results on test videos show that the proposed algorithm has good detection effect,it is a high efficient flame detection algorithm.
作者
熊昊
李伟
XIONG Hao;LI Wei(Naval Aviation University,Yantai 264001,China)
出处
《传感器与微系统》
CSCD
2020年第1期143-145,149,共4页
Transducer and Microsystem Technologies
关键词
火灾检测
特征提取
支持向量机
fire detection
feature extraction
support vector machine(SVM)
