基于视频图像的火焰检测.pdfVIP

Computer Science and Application 计算机科学与应用, 2016, 6(3), 171-177 Published Online March 2016 in Hans. /journal/csa /10.12677/csa.2016.63021 Flame Detection Based on Video Kai Li, Shengbo Li, Rui Liu, Jie Wang, Dan Liu Faculty of Computer Engineering, Huaiyin Institute of Technology, Huaian Jiangsu th th th Received: Mar. 7 , 2016; accepted: Mar. 26 , 2016; published: Mar. 29 , 2016 Copyright © 2016 by authors and Hans Publishers Inc. This work is licensed under the Creative Commons Attribution International License (CC BY). /licenses/by/4.0/ Abstract Fire as a natural disaster is very common in our life. Therefore, flame detection timely is very im- portant for the prevention of fire. In this paper, a novel fire detection method is proposed based on background subtraction, color feature and shape features. Firstly, the moving object is ex- tracted with background image difference based on the dynamic feature of fire. Then, the target areas are extracted exactly by five rules of color feature and three shape features of fire. Finally, the suspected flame is detected. The proposed method shows more effective for fire detection and presents high use value. Keywords Fire Detection, Background Subtraction, Color Feature, Shape Feature 基于视频图像的火焰检测 李 凯，李生波，刘 瑞，王 杰，刘 丹 淮阴工学院计算机工程学院，江苏 淮安 收稿日期：2016年3月7 日；录用日期：2016年3月26 日；发布日期：2016年3月29 日 摘 要 火灾是当今生活中十分常见的自然灾害，因此及时检测火焰对预防火灾的发生十分重要。本文提出了一 文章引用: 李凯, 李生波, 刘瑞, 王杰, 刘丹. 基于视频图像的火焰检测[J]. 计算机科学与应用, 2016, 6(3): 171-177. /10.12677/csa.2016.63021 李凯 等 种基于背景差分和颜色、形状特征的火焰检测方法。该算法首先采用背景差分法，根据火焰燃烧的动态 性实现对运动目标的提取，再根据火焰颜色特征的五条规则以及火焰的三个形状特征对目标区域进一步 提取，得到最终的检测结果。该方法火焰检测率较高，具有较好的实用价值。 关键词 火焰检测，背景差分，颜色特征，形状