摘要
树枝叶等植物的晃动是造成户外监控背景复杂性的主要因素。该文提出一种基于视频分析的运动对象检测策略，利用快慢双背景模型检测运动变化像素，并通过邻域像素的相关性检测消除虚假运动对象。实验表明此算法对户外监控不但速度快而且检测效果理想，满足实用化要求。
Swaying trees frequently cause false alarms for outdoor surveillances.A practical scheme of detecting moving objects based on video analysis is proposed in this paper,which uses fast and slow background models to detect changed pixels,and eliminates false moving objects by checking the correlations among neighboring pixels. Experiments show the scheme of detecting moving objects is fast and robust in outdoor scenes.
出处
《计算机工程》
CAS
CSCD
北大核心
2003年第16期16-17,121,共3页
Computer Engineering
基金
国家自然科学基金资助项目(60075006)
关键词
视频监控
背景减除法
运动对象
Video surveillance
Background subtraction
Moving objects
