基于纹理相关性的飞机货舱视频烟雾检测方法

Video smoke detection in aircraft cargo hold based on texture correlation

提出一种图像型烟雾探测方法并进行实验验证。采用主动红外CCD摄像机获取图像,利用四帧差分法提取烟雾可疑区域,通过计算图像帧序列的纹理相关性特征值来设定烟雾阈值,以判断是否有火灾发生。在模拟飞机货舱内进行真实火源和干扰火源实验,实验结果表明,烟雾和干扰项可以通过纹理相关性特征值进行区分。该方法在满足运输类飞机适航标准要求的同时,提高了烟雾检测速度,为视频烟雾探测技术在飞机货舱中应用提供研究基础。

An image smoke detection method was put forward and experimentally verified. Using lhe aclive infrared CCD camera to obtain the image, the four-frame difference method was used to ex tract the smoke suspicious area, The smoke threshold was set by calculating the lexture characteristic value of the image frame se quence to judge whether a fire occurred. Experiments of real fire source and disturbing fire source were carried out in simulated air crafl warehouse. Experimental results showed lhat smoke and in terference items can be distinguished by texture retated eigenvalues. The method meets the requirements of airworthhless of trans porl aircraft and improves the detection speed of smoke, which provides research basis for the application of video smoke dcteetio technology in aircraft cargo hold.