一种稳健的运动探测与背景维护算

针对家庭安全监控需要，提出一种稳健的运动对象探测算法。首先给出三帧差运动对象检测算法，同时利用像素的亮度分量和色度分量避免由于大尺寸对象闯入引起图像背景像素变化造成运动对象检测的失败。然后提出一种三关键背景维护算法，用于清除背景振动像素，减少虚假报警。实验表明，这种运动探测算法对视频监控系统消除虚假异常报警具有良好的通用性，实时性和稳健性。

A Robust Algorithm for Moving Object Detection and Background Maintenance

This paper presents a robust algorithm for detecting moving objects in video security surveillance, which is composed of three frame difference and three key value. In order to solve the problem that a big breaking in object causes the background to change its illumination and make detection results invalid, three frame difference is used to detect moving object. It uses the illumination and the chroma in the YUV color space to find moving objects. Three key value is a new method for background maintenance, which can eliminate the swaying objects in the background and minimize false alarms. The experiments in the house security surveillance show that this algorithm is real time and robust.