Research And Implementation Of Adaptive Surveillance System Based On Face Recognition

In some public places,people flow is relatively large,which brings difficulties to security work.In order to solve this problem,we can use the face recognition technology to deal with the captured video.Then,the facial features are saved and the number of features are counted.The face whose frequency is higher than the threshold is marked out.According to the result,the security guard is reminded to pay more attention to the man.An adaptive monitoring system based on face recognition is designed.The system is mainly applied in the filed of public security,and its main function is to realize intelligent monitoring and real-time monitoring without the deliberate management of security personnel.It consists of 4 parts,namely video image acquisition part,face detection part,face recognition part and warning part.In the acquisition part of video image,video frames are retrieved and denoised by calling the OpenCV library function.In the face detection part,the Adaboost algorithm base on haar-like features is studied.Face detection simulation experiments are carried out though face database of CUM-PIE,network pictures and video images.The test result meet the design requirements of the system.In the face recognition part,the SIFT algorithm is studied.Face recognition simulation experiments are carried out though video face database of FRGC v2.0 and the optimal parameters are selected.The test result meet the design requirements of the system.The expected goal of the system is to be able to automatically detect and recognize the face of video image,and to count the frequency of the same face.It can set the threshold size to achieve the warning effect.The research and system design of face recognition technology are completed.After the test,the system runs stably,has a friendly interface and good real-time performance,and has a good effect of face detection.The test results demonstrate the feasibility of the system.