基于混合高斯模型及小波分析的烟雾特征提取分析

Smoke Character Extracting and Analyzing Based on Mixture Gaussian Model and Wavelet Analysis

烟雾探测在日常生产生活中具有重要意义,是预见火灾隐患的重要手段。基于小波分析的图像型火灾烟雾特征提取的方法,采用混合高斯模型对背景进行建模,从而提取可疑的运动区域。相对于传统的运动提取方法,此方法能够有效的抑制光线变化以及噪声带来的干扰,实验结果表明,此判据具有较高的准确性。

A smoke detection method was proposed based on methods such as mixture Gaussian model and wavelet analysis etc.Mixture Gaussian model was used in the process of extracting suspicious area.This method could restrain the disturbance caused by light variety and other noises,thus,the motional area could be clearer and more integrity.The experiment result shows that this criterion has a high veracity.