一种基于CCH的运动目标快速检测算法  

A Fast Algorithm for Moving Object Detection Based on CCH Descriptor

作　　者：裴冠军[1] 周荷琴[1] 

机构地区：[1]中国科学技术大学自动化系

出　　处：《电子技术（上海）》2011年第1期7-9,共3页Electronic Technology

摘　　要：运动目标的检测是智能监控系统和视频理解标注中的一个底层处理环节,这一环节的性能直接影响到中高层目标本体特征和运动特征的提取、目标分类、目标跟踪以及目标异常行为的分析。针对一些监控场景下常用算法速度较慢、对阴影过于敏感等缺陷,提出一种新的区域内容对比度直方图来描述小区域的特征,并利用该特征建立混合直方图背景模型,再通过减背景实现运动目标的快速检测。实验结果表明算法快速有效。Moving object detection is an essential link in intelligent video surveillance system or in video understanding and labeling. It will directly affect the subsequent upper processing tasks such as the abstraction of object features and movement parameters, the object classification, object tracking, abnormal activity detection, etc. To improve the detection speed using conventional algorithms which work slowly in some big size videos and be easily disturbed by shadows and illumination, a novel block feature descriptor is introduced, and based on this descriptor a mixture histogram-background model is constructed. Then through background subtraction, the moving foreground is detected. Experiments demonstrate this algorithm is fast and efficient.

关 键 词：视频监控 背景模型 运动检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]