基于小波的实时烟雾检测*

传统的离子式、吸气式、光电式等烟雾检测器在大空间中检测烟雾时,会受到发射信号与接收信号之间的距离、平面角度、精确对准等限制,无法对整个空间的烟雾状况进行描述.新的方法通过监测区域的摄像机拍摄的视频图像序列,进行小波变换,分析图像帧在时域和空域的频率特性,来确定被监测区域是否有火灾烟雾的发生.实验证明该方法不受空间高度、热障、易爆、有毒等环境的限制,并且有灵敏度高、抗干扰力强、适用范围广等特点.

Wavelet Based Real time Smoke Detection

The traditional ionization smoke detector, aspirated smoke detector and photoelectric smoke detector etc were limited by distance between the emissive signal and received signal ,plane angle and precisely aims when it examined the smoke in large space, unable to carry on the detection to the entire spatial smoke condition. The new method was analyzed the temporal frequency and the spatial frequency characteristic in video frames through the video sequence of the camera monitoring the scene were wavelet transforms, which determined the region whether have the fire smoke occurrence or not. The experimental results were proved that the method was not limited by the large space, the thermal barrier, explosive, virulent and so on, and it has the characteristic of high sensitivity, the strong anti-disturb, the broad applicable scope and so on.