改进的基于统计模型的前景检测方法  

Improved foreground detection based on statistical model

作　　者：强振平[1] 刘辉[2] 尚振宏[2] 陈旭[1] 

机构地区：[1]西南林业大学计算机与信息学院,昆明650224 [2]昆明理工大学信息与自动化学院,昆明650051

出　　处：《计算机应用》2013年第6期1682-1685,1694,共5页journal of Computer Applications

基　　金：云南省科技计划项目(2009CA013)

摘　　要：针对基于统计模型的前景检测方法进行改进:一方面,背景模型中记录特征向量属于背景的历史最大概率,在当前帧像素点特征向量与背景模型中已有特征向量匹配时,利用历史最大概率提高其更新速度,使其尽快融入背景;另一方面,对利用贝叶斯决策规则检测的前景目标,剔除其轮廓信息后与背景的空间特征进行匹配,减少阴影对前景检测的影响。实验结果表明,与MoG方法和Li的统计模型方法的前景检测相比,该方法在阴影剔除以及大目标物体遮挡背景恢复等方面都有明显改进。In this paper, the main idea was to improve the foreground detection method based on statistical model. On one hand, historical maximum probability of which feature vector belongs to background was recorded in the background model, which could improve the matched vector＇s updating speed and make it blended into the background quickly. On the other hand, a method using spatial feature match was proposed to reduce the shadow effect in the foreground detection. The experimental results show that, compared with the MoG method and Li＇s statistical model method, the method proposed in this paper has obvious improvement in shadow remove and obscured background restoration of big target object.

关 键 词：前景检测 背景模型 统计模型 贝叶斯决策 阴影剔除 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]