视频火灾自动预警中烟雾检测的研究

首先分析了在火灾发生时烟雾的静态特征和动态特征，主要包括烟雾的颜色特征、形状不规则性和扩散性。针对烟雾的颜色特征、形状不规则性及扩散性设计了相应的检测算法，实现了对视频中的烟雾的检测并自动报警。实验结果表明，结合烟雾的静态特征和动态特征设计的检测算法具有较好的识别效果。

Smoke Detection Studying in Automatic Fire Early Warning Based on Vedio

First,analysis the smoke's static characteristics and dynamic characteristics when the fire broke out at,including of smoke color,irregular shape and spreading. For this characteristics of smoke,designed the corresponding detection algo-rithm,implemented the smoke detection and automatic alarming. Experimental results show that the detection algorithm based on the combination of the static and the dynamic characteristics has better recognition effect.