门禁监控视频软件设计与实现
摘要:传统的门禁视频监控系统一般是无法对人员进出门禁时进行人头检测,只能通过简单的硬件设施将监控画面传到后端,由监控人员人工进行监视判断。这样会加大监控成本也增加了监控人员的压力,同时也对监控人员的身体健康产生较大影响。所以就有必要设计一款新型门禁监控视频系统来实现人头检测功能,本文设计出的一款基于WPF的门禁监控视频系统,将基于Haar特征的AdaBoost人头检测算法和基于Hog特征的SVM人头检测算法进行比较,在人员进出门禁系统时进行人头检测并将检测结果以图像方式进行保存。
基于WPF的门禁视频监控系统,采用C#与EmguCV的技术,是实现对人员出入门禁系统时进行人头检测。论文主要分析了两种算法的特征值的提取方法以及利用这两种算法进行人头检测,设计出一个可以检测人头的门禁监控视频。
Access control monitoring of video software Design and Implementation
Abstract:Traditional access control video surveillance systems are generally unable to detect the head of personnel entering and leaving access only through a simple hardware facilities will spread to the rear end of the monitor screen, the monitoring personnel to monitor artificial judgment. This will increase monitoring costs are increasing pressure monitoring personnel, it will be have a greater impact on health monitoring personnel of the body. Therefore it is necessary to design a new access control system to achieve video head detection, designed out of a WPF-based access control and video systems, will head detection algorithm based on AdaBoost and Haar features characteristic of SVM-based Hog head detection algorithm comparing head detection and test results performed at the personnel access to the access control system to save the image mode..
WPF-based access control video surveillance systems, using C # and EmguCV technology is to achieve when people head out access system testing. Thesis analyzes the eigenvalues extraction and the use of these two algorithms two algorithms detect head, designed to detect a surveillance video access control head.
目 录