视频火焰检测综述

Abstract

基于传感器的传统火灾检测系统已经不能满足实际需求。随着计算机技术和数字图像处理技术的发展，视频火焰检测作为一种新型有效的早期火灾探测技术，已经受到人们的广泛关注。本文介绍了视频火焰检测流程，着重分析火焰的图像特征，包括基于单帧的静态特征和基于多帧的动态特征，并探讨了典型的特征提取算法，对多特征融合算法进行了分类比较，最后展望了视频火焰检测的发展趋势。 The traditional fire detection system is an unsatisfactory way of detecting fire based on some sensors. As an effective type of early fire detection technology, video fire detection has received extensive attention recently with the improvement of technology of computer and digital image process. The process of video fire detection is shown here. The flame image characters are discussed such as the static characteristics in single frame and dynamic characteristics in multiple frames. Some typical methods of fire image feature extraction are presented. Then the fusion algorithm with multiple features is classified and summarized. Finally, the development of video fire detection is prospected.