Research And Design Of Face Recognition Intelligent Access Control System

With the development of economy and technology, people put forward higher requirements for security. As one of the advanced biometric technology, face recognition is gradually applied in our lives. The combination of the face recognition with access control would have a promising future. But the common face recognition access control systems often choose PC as the operating platform of system, which brings the disadvantages of a high cost and tedious installation. In this thesis,according to the disadvantages of common access control systems, the face recognition technology and embedded technology are combined to research and design the new access control system based on face recognition.This thesis firstly analyzes the requirements of the access control system, then puts forward the technical indicators of new access control according to the actual application requirements, and completes the design of the system. The system selects the Smart210 as the main control unit, the Linux operating system as the software design platform, and uses Video4Linux2 framework to realize the real-time acquisition of image. In order to remove the interference factors in the image, the preprocessing steps including de-noising, grayscale, and histogram equalization are performed before the face recognition. This thesis simulates and tests common face recognition algorithm and focuses on contrasts recognition rate and recognition time-consuming, and then selects the face recognition algorithm for the system after research. A theoretical analysis of the Adaboost algorithm and PCA algorithm is conducted, then the Adaboost algorithm achieve the functions of face detection , the classifier is trained for blink detection to determine the real face, excluding the case when the face picture is used to open the door. In order to improve the anti-interference of the system to light,the face recognition algorithm based on block LBP operator and PCA is proposed.Then,according to the disadvantage of the low rate of PCA algorithm in recognizing the side face,the front face reconstruction method is proposed based on auto encoder. By using MATLAB for the simulation,the front face is reconstructed,and then PCA algorithm is combined to test the results of the simulation. The QT framework and the OpenCV computer vision library are used to complete the design of human-machine interface and system identification module,and the VS2013 and MySQL database are used to complete the design of the information management center on the Windows platform, mainly achieving the management of administrator and legal personnel information.The thesis completes the whole design of the system,and tests the real-time and recognition rate of the face recognition intelligent access control system. The results of test show that the system can perform the function of face recognition well, and the real-time and the recognition rate can meet the requirements. The intelligent access control system researched and designed in this thesis uses the face as an identity to open the door,and conducts an unified and centralized management of information of the visitors. The technology involved has research value, and the system is easy to use and has a broad application market.