基于背景动态更新与暗通道先验的火灾烟雾检测算法  

Based on dynamic background update and dark channel prior of fire smoke detection algorithm

作　　者：赵亮[1] 骆炎民[1,2] 骆翔宇[1] 

机构地区：[1]华侨大学计算机科学与技术学院,福建厦门361021 [2]华侨大学机电及自动化学院,福建厦门361021

出　　处：《计算机应用研究》2017年第3期957-960,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(61170028);华侨大学科研启动项目(14BS215);华侨大学研究生科研创新能力培育计划资助项目

摘　　要：烟雾检测在现代智能消防中有着重要的应用前景,随着计算机视觉和模式识别技术的发展,基于视频图像的火灾烟雾检测算法不断被提出。针对目前检测方法适应性不强、在复杂环境下检测性能不高的问题,提出了一种基于背景动态更新和暗通道先验的烟雾检测算法。算法首先通过改进的背景动态更新算法提取运动前景;然后,结合暗通道先验知识确定前景中的疑似烟雾区域;最后,利用烟雾颜色特征、旋转不变的LBP纹理特征和HOG特征的线性融合通过最近邻分类器(KNN)进行识别。通过在多个视频场景下的实验,表明该算法受环境因素影响较小,且具有良好的烟雾检测能力。Smoke detection has important application prospects in the modern intelligent fire,with the development of computer vision and pattern recognition technology,the fire smoke detection algorithm based on the video image constantly being proposed. For the current detection methods are not flexible and detection performance is not high in a complex environment,this paper proposed a fire smoke detection algorithm based on dynamic background update and dark channel prior. Firstly,it improved the dynamic background updating algorithm for extracting moving foreground. Then,combined with the dark channel prior knowledge to determine the foreground of suspected smoke area. Finally,it used linear combination of smoke color features,rotation invariant LBP texture feature and HOG features by the KNN classifier for identification. Through the experiments on more than one video scene,it shows that the algorithm is less affected by environmental factors,and has good capability of smoke detection.

关 键 词：背景估计 暗通道先验 颜色特征 旋转不变的LBP HOG特征 KNN分类器 烟雾检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]