特定区域实时监测系统设计与实现

发布时间：2018-06-05 20:51

本文选题：安全问题 + 智能视频监控技术　；参考：《安徽工业大学》2017年硕士论文

【摘要】：安全问题是现今社会一同关心的重要事件,所以安检是各行各业,特别是有着潜在危险的高风险行业所必须重视的一项任务,并在近年来处于越来越重要的位置。时间的推移以及应用的增加,使得传统人工监控的缺陷变得越来越突显。因此作为现代安检的主要手段之一,智能视频监控技术有着非常良好的发展前景。智能视频监控技术(intelligent video surveillance)起源于计算机视觉(computer vision)的技术,其对视频图像进行解析,获取视频中的有效信息,进而能够在某些方面帮助甚至取代人工监控。因此在本文中,我们将利用图像处理结合智能视频监控技术解决特定区域下(本文取为服务器机房)的一些问题。首先,在进入服务器机房之前必须先进行人员监控。服务器机房工作人员及某些不是工作人员但被准予进入的人员可以入内,其他人不得入内。其次,进入服务器机房之后,需要进行人物越界实时监测和头部运动逾时监测。第一部分为人物越界实时监测:能够实现对人物进行近距离持续动态跟踪,并可对视频中特定矩形区域进行越界实时监测,一旦出现越界行为,立即进行报警。第二部分为头部运动逾时监测:能够定位到人物头部并实现对人物进行头部运动逾时监测,一旦头部转动时间达到最长时间,立即进行报警。本文在视频监控中使用人脸识别的技术可以使人员进出服务器机房的安全问题得到改善。系统中的人物越界实时监测和头部运动逾时监测具有很重要的意义,不但能为个人或者单位节省人力成本,而且能在整日的实时监控中完成预警功能。本文搭建的特定区域实时监测系统有着很好的二次开发接口,以便于其后续研发工作的迅速进行。
[Abstract]:Safety is an important issue that the society is concerned about together, so security inspection is a task that must be paid attention to by all walks of life, especially the potentially dangerous high-risk industries, and is in a more and more important position in recent years. With the passage of time and the increase of application, the defects of traditional manual monitoring become more and more obvious. Therefore, as one of the main means of modern security inspection, intelligent video surveillance technology has a very good development prospects. Intelligent video surveillance technology originates from computer vision technology, which parses video image, obtains effective information in video, and can help or even replace manual surveillance in some aspects. Therefore, in this paper, we will use image processing and intelligent video surveillance technology to solve some problems in a specific area (server room in this paper). First, people must be monitored before entering the server room. Server room staff and certain non-staff but admitted personnel may enter, others may not enter. Secondly, after entering the server room, it is necessary to carry out real-time monitoring and head movement delay monitoring. The first part is the real-time monitoring of the character crossing the boundary: it can realize the close and continuous dynamic tracking of the character, and can monitor the specific rectangular area in the video in real time. The second part is the head movement overtime monitoring: it can locate the head of the character and realize the head motion overtime monitoring. Once the head rotation time reaches the longest time, the head can be alerted immediately. In this paper, the use of face recognition in video surveillance can improve the security of people entering and leaving the server room. It is of great significance that the real time monitoring and the head movement time-out monitoring of the characters in the system can not only save the manpower cost for individuals or units, but also complete the early warning function in the real-time monitoring of the whole day. The real-time monitoring system in this paper has a good secondary development interface, so as to facilitate the rapid development of its follow-up work.
【学位授予单位】：安徽工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TN948.6

【参考文献】