摘要
为了实现对视频序列中运动目标的实时检测,提出了一种新的运动目标检测方法.该方法对相邻两帧进行差分处理,并通过对帧差结果进行方差分析构建背景模型;用方差作为阈值将差分图像划分为运动变化区域和背景区域;将运动区域与构建的背景模型拟合,区分出运动目标和显露区域,并将显露区域收入背景,实现背景的更新.实验结果表明,该方法运算速度快、准确率高,能够对运动目标进行实时检测.
A new method on moving objects detecting was presented for the moving objects real-time detection on the video sequence.The subtraction result was obtained through the adjacent frames subtraction.The background model was built according to variance analysis on the subtraction result.Then with the threshold that was the same as variance,the subtraction result was separated into two parts： moving area and background area.At last,moving area was compared to background to locate the moving objects and the uncovering area.The background was updated with the uncovering area continuously.It could achieve clear moving objects,and can meet the need of the real time video surveillance and monitoring system with a high operation speed and accuracy.
出处
《郑州轻工业学院学报（自然科学版）》
CAS
2010年第6期91-94,共4页
Journal of Zhengzhou University of Light Industry:Natural Science
基金
国家科技支撑计划项目(2006BAK01A38)
河南省教育厅自然科学研究计划项目(2009A510018)
郑州轻工业学院博士基金项目(2007BSJJ005)
关键词
方差分析
运动目标实时检测
帧间差分
阴影检测
背景减法
variance analysis
moving objects real-time detection
frames difference
shadow detection
background subtraction
