基于混合高斯和均值滤波法的运动检测方法

Moving Object Detection Based on Mixture Gaussian Model and Median Filter Method

针对智能视频处理技术中运动目标检测问题,提出了运动目标检测中背景动态建模的算法。该算法有效地结合了均值滤波法和混合高斯模型法的优点,不仅提高了系统对快速运动目标的检测能力,而且消除了单独使用混合高斯法时容易产生的"假"前景区域现象,同时提高了前景检测率。结果表明,在有诸多不确定性因素的序列视频中构建的背景具有较好的自适应性,能迅速响应实际场景的变化。

Aiming at the problem of detecting moving objects in intelligent video processing technique,an improved moving object detecting method based on the mixture Gaussian model and median filter was presented in this article.This algorithm has the advantages of both median filter method and mixture Gaussian model method.Not only has it enhanced the ability of detecting quick moving objects,but also eliminate fake foreground in which often resulted by the median filter method.The experimental results indicated that the model had preferable adaptive performance to the scene with many uncertain factors,and it corresponded quickly.