基于混合高斯运动检测模型与多特征的烟雾识别算法

Smoke Detection Algorithm Based on Gaussian Mixture Motion Detection Model and Multiple Features

为了有效克服复杂环境下动态、静态疑似烟雾物体的干扰,实现对烟雾的实时准确检测,提出基于混合高斯运动检测模型与多特征分析的烟雾识别算法。首先应用优化的混合高斯运动分析模型对视频图像序列进行运动区域提取,然后依据烟雾的颜色特性、形状不规则性及面积扩散特点,对提取的疑似烟雾运动区域进行分析与筛选,从而判定出其是否为烟雾。实验结果表明:该算法可实时提取视频中的烟雾区域,并有效剔除疑似烟雾区域的干扰,具有良好的烟雾识别能力。

In order to overcome interference with the similar static or dynamic smoke in complicated environment and realize real-time and accurate smoke detection, a smoke detection algorithm based on Gaussian mixture model and multiple features is presented. First, the motion region of video is obtained by modified Gaussian mixture model, and is analyzed according to color,irregular shape and diffusion features of the smoke. Then the conclusion is acquired that the motion region is a smoke region or not.The experimental results show that the algorithm has favorable smoke detection performance, which can acquire real-time smoke information in video and eliminate the interference of similar smoke area.