摘要
在室内视频监控系统中,光照条件对于运动物体的检测与分类将产生显著影响。本文归类分析了室内照明条件对于运动目标检测算法的不同影响,提出一种能够快速地适应光照条件变化的运动目标检测算法。该算法结合了帧间差分与背景模型算法的优点,并且引入亮度信息进行前序的处理;因此能够快速准确地检测出运动目标,同时算法对不同光照条件具有较高鲁棒性。通过相关仿真实验,证明了该算法的实时性与可靠性。
In indoor video surveillance systems, the disturbance of illumination conditions can influence the detection and category of moving objects greatly. This paper analyzed the influence of different illumination conditions to the moving-object detection in detail. It also proposed a new moving-object detection algorithm that could adapt illumination changes rapidly. With the illumination information processing as former procedure, this algorithm combined the merits of background subtraction method and frame differencing method. Thus it can detect the moving objects with both rapid and accuracy, at the same time, it has much robust to different illumination conditions. The experimental results verified the real-time and reliability of the algorithm.
出处
《照明工程学报》
2007年第3期12-15,22,共5页
China Illuminating Engineering Journal
基金
湖南省教育厅资助科研项目(05C411)
湖南师范大学博士科研启动基金资助项目(I050602)
关键词
运动目标检测
光照变化
视频监控
混合高斯模型
moving-object detection
illumination change
video surveillance
mixture of Guassian （MoG）
