摘要
烟焰作为火灾最为明显的标志,对于其进行实时检测,在火灾监控方面有着重要的意义。传统的离子式、光电式、吸烟式等烟雾探测器受到空间和时间的限制,不能进行实时全面的探测[1]。本文提出了一种基于视频的烟焰检测算法,采用帧差法提取出差值图像,经过图像处理,获得了理想的烟焰轮廓,根据轮廓大小来确定被监测区域是否有火灾的发生。实验证明,该方法实时准确、灵敏度高、抗干扰能力强、适用范围广,具有广阔的应用前景。
The real-time detection of smoke and flame which is the most obvious sign of fire disaster is very important in the fire monitoring. Being subject to space and time , traditional smoke detectors, such as ion-type, photoelectric type, smoking type, can＇t complete fully real-time detection. In this paper, a smoke and flame detection algorithm based on video is presented, which extracts the difference images by using frame difference method. After image processing, an ideal profile of smoke and flame can be gotten. According to the contour area, it can be determined whether there is a fire disaster happened.Experiments show that the method has advantages in high accuracy , real-time, sensitivity, anti-interference ability, wide application and good application prospects.
出处
《传感器世界》
2010年第10期14-17,共4页
Sensor World
基金
山西省自然基金项目资助(编号:20070120003)
关键词
烟焰
图像分割
视频
帧差法
smoke and flame
image segmentation
video
frame difference method
