基于多特征融合的户外视频烟雾检测算法
Outdoor Video Smoke Detection Algorithm Based on Multi-Feature Fusion
王习越 1陶青川1
作者信息
- 1. 四川大学电子信息学院,成都 610064
- 折叠
摘要
针对目前户外视频烟雾检测的检测率低、误检率高的情况,提出一种新的基于多特征融合的户外视频烟雾检测的算法.在YUV空间下采用码书模型进行背景建模,提取出疑似烟雾区域;对疑似区域进行烟雾的颜色、模糊性、形状以及区域中心稳定性检测;将各个特征的检测结果进行贝叶斯分类器融合,并用于判断烟雾是否存在.实验表明,该检测算法鲁棒性高、能够准确地检测出视频图像中出现的烟雾,具有一定的实用性.
Abstract
Proposes new multiple feature fusion based smoke detection algorithm,which deals with situations like low detection rate and wrong detection in outdoor videos.Uses code book in YUV for background modeling to extract smoke area.Detects the color,shape,and center position of the suspicious smoke area.Carries out the fusion on the detect outcome using Bayesian classifiers.Takes the fusion result as a symbol to judge whether the smoke actually exist.Experiments show that the proposed algorithm is robust and accurate in smoke detection in video images.
关键词
本模型/多特征融合/贝叶斯分类/烟雾检测Key words
Code Book/Feature Fusion/Bayesian Classification/Somke Detection引用本文复制引用
出版年
2015
NETL
NSTL
维普
万方数据