Font Size: a A A

Video Surveillance, Face Detection And Recognition

Posted on:2011-05-28Degree:MasterType:Thesis
Country:ChinaCandidate:P F GuoFull Text:PDF
GTID:2208360302498278Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Existing video surveillance systems in many residential district has an application, but as individual home users, which itself has a major weakness as a family with a monitoring system, video surveillance storage possesses too large,which will affect the monitoring cost. The more advanced infrared thermal imaging technology for monitoring is expensive for home users,which costs too much. This essay will discuss a new control measure, using face recognition technology, and establish a feature picture library, through the camera, on the visitor's information extraction determine the identity of visitors. The advantage is the capacity to save, just face image, and can be reused. Accumulated over time, features will fullfill picture library data, the system will determine the accuracy more and more.When the system determines visitors to acquaintances (in safety case), do not produce alarm, otherwise select and create alert.Based on the research of other relevant literature, an improved method to update background based on background difference method and adjacent frame difference method is given in this essay, analysis both static and dynamic cases, the number of non-zero pixel differences, by setting the division value to achieve background update.,in the condition of setting different value of division, in laboratory conditions(lm-5m area before the monitor camera,30-megapixel camera)Adaboost algorithm using video surveillance to measure face which comes into monitoring area, and had normalized the size of the image. For the detected face images, has established my own laboratories face images, and using principal component analysis of PCA method to achieve face recognition. The alarm settings, this paper presents two ways.Based on the above study of video surveillance systems, video surveillance system can detect and recognize the face of charactors coming into monitoring area, while achieving the background of the update, in the laboratory environment, it has a good experimental results.
Keywords/Search Tags:OpenCV, background update, face detect, face recognition, alarm setting
PDF Full Text Request
Related items