基于贝叶斯模型的动态背景检测  

Bayesian Modeling for Dynamic Background Detection

作　　者：刘英霞[1] 贺长伟[1] 王欣[1] 

机构地区：[1]山东大学信息科学与工程学院,山东济南250014

出　　处：《系统仿真学报》2007年第21期5042-5045,5058,共5页Journal of System Simulation

基　　金：国家自然科学基金(60172022)

摘　　要：利用贝叶斯模型结合核密度函数理论,给出了在复杂的动态背景下检测出运动目标的方法。首先,在图像相关性的基础上,利用非参数核密度估计的方法,建立前景和背景的核密度函数,再利用贝叶斯理论,估计出背景和前景的先验概率,两者相结合得到一个估计的阈值,从而实现目标和背景的分类。该方法对阈值的选取进行了理论分析,实验证明,提出的方法能够在动态背景中准确地检测出目标。Bayesian model and kernel density function were applied to detect the moving object from the dynamic background. On the basis of the image correlation and the nonparametric kernel density estimation, kernel density functions of background and foreground were built., Bayesian theory was used to estimate the prior probabilities of the foreground and the background, so an estimative threshold was obtained and the classification of the background and the foreground was realized. The method gave a theoretical analysis of the threshold selection. The result of the experiment is performed that the method can detect the object from the dynamic correctly.

关 键 词：非参数核密度估计 贝叶斯模型 动态背景 阈值选取 目标检测 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]