低浓度烟雾检测方法研究与仿真

摘要

研究烟雾准确检测的问题,由于烟雾扩散和泄漏,空气浓度的变化,传统烟雾检测方法,必须要保证烟雾达到一定浓度,与外界环境的像素差异达到一定程度之后才能进行检测.当烟雾泄漏之后快速扩散,浓度被迅速稀释,烟雾像素与外界差异达不到检测要求,造成传统方法不能准确检测烟雾泄漏.为解决上述问题,提出一种新的烟雾检测方法.利用图像颜色模型方法,结合数学形态学提取出初始目标,通过对大量烟雾的调查研究,找出烟雾在特定颜色空间中的分布,进而进行有效的检测,克服了传统方法的弊端.实验证明,改进方法实现了低浓度烟雾的准确检测,取得了满意的效果.%In traditional smoke detection method, the difference between smoke pixels and background pixels must be obvious enough. This paper put forward an image detection method. Using image color model and combining math-ematical morphology, the initial goals were extracted. Through the investigation and study for a lot of smokes, the smoke distribution in specific color space was found out, and then the image was identified. Experiments show that the method realizes the automatic detection of low concentration smoke, and achieves satisfactory results.