摘要
为了快速准确地对视频序列中的遗留物体进行监控,对传统的双背景遗留物检测方法进行了对比改进,提出改进后的快慢混合高斯背景模型算法,对运动区域进行检测。利用改进后的Camshift的跟踪算法对运动目标进行跟踪并且基于图像信息熵的概念对运动目标进行提取。实验结果表明,该算法能够准确检测出场景中的遗留物体,有效提取运动目标并显示遗留物主,具有较强的鲁棒性。
To detect the abandoned object quickly and accurately in the video sequences,the abandoned object detection based on dual-background model method was improved.A method called the improved dual-background model method was presented to extract movement regions,then movement regions were tracked based on Camshift method and a key frame of movement regions was extracted based on entropy of image information.Experimental result showed that the algorithm could detect the area of abandoned object and extract movement regions accurately,and had strong robustness.
出处
《计算机工程与设计》
CSCD
北大核心
2014年第6期2192-2196,共5页
Computer Engineering and Design
基金
重庆市科委自然科学基金项目(2010BB2399)
关键词
智能视频监控
遗留物检测
双背景模型
CAMSHIFT
图像信息熵
intelligent video surveillance
abandoned object detection
dual-background
Camshift
entropy of image information
