背景差分与三帧差分结合的运动目标检测算法  

A Method for Moving Object Detection Based on Background Subtraction and Three- Frame Differencing

作　　者：卢章平[1] 孔德飞[1] 李小蕾[1] 王军伟[1] 

机构地区：[1]江苏大学机械工程学院,江苏镇江212013

出　　处：《计算机测量与控制》2013年第12期3315-3318,共4页Computer Measurement &Control

基　　金：高等学校博士学科点专项科研基金(20113227110007)

摘　　要：针对基于混合高斯模型背景差分法对光照突变敏感的问题,提出了背景差分法与三帧间差分法相结合的运动目标检测算法;首先利用当前帧与混合高斯模型建立的背景模型差分,快速检测出运动变化区域;然后,通过与设定的阈值比较,判断场景内是否发生光照突变,场景中若未发生光照突变,则采用混合高斯模型背景差分法提取运动目标,若发生光照突变,采用三帧间差分法提取运动目标;实验结果表明,在光线发生突变情况下,文中提出的算法同样能够取得较好的检测效果,具有很强的适应性和鲁棒性,可用于实时监测系统。The background subtraction method based on Gaussian mixture model is sensitive to light mutation,a combining method of background subtraction and three-frame differencing is proposed to make up its deficiency.First,the current image minuses the background model to detect the moving area rapidly,then compared to the threshold value set previously to judge whether light mutation happens in the monitoring area.If there is no light mutation,the method of the background subtraction based on Gaussian mixture model is selected to extract moving objects;in the situation of light mutation,the method of the three-frame differencing is selected to extract moving objects.At last,the experiment shows that the proposed method can also have a better detection result when light mutation happens.Meanwhile the robustness and suitability are also acceptable,and it can be applied into real-time video monitoring system.

关 键 词：光照突变 混合高斯模型 背景差分法 三帧差分法 实时监测系统 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]