Design And Implementation Of Intelligent Access Control System Based On Face Recognition

With the continuous development of information technology and the increasing progress of modern society,at the same time,with the hidden dangers of unsafe factors,the intelligent security system has been popularized.Access control system,as the core part of security system,has increasingly become the focus of attention from all walks of life.The traditional access control system using cards and passwords is unable to meet people’s growing demand for safety because of its shortcomings such as single function and easy loss.Because of its high uniqueness and stability,biometric identification technology is very suitable for the security system,and it has become a hot research direction in the entrance guard system.In this paper,a set of intelligent access control system based on face recognition is designed for small office space.The system consists of two parts,embedded access control end and face recognition server,which have the function of authority control and attendance.The main work of the full text is as follows:1.Face recognition theoryAccording to the face recognition system designed in this paper,a suitable method is needed.After the experiment,the classical face detection algorithm AdaBoost and the recently popular MTCNN face detection algorithm improved by the convolution neural network are compared,and the MTCNN face detection algorithm is selected.On the basis of the in-depth study of the mainstream face recognition algorithm,a face recognition algorithm based on LC-KSVD-CRC is proposed.The theoretical analysis results show that the algorithm has a high recognition rate.But the performance is not as good as FACENET in the actual scenario test,so FACENET face recognition algorithm is chosen.2.Design of intelligent door control systemThe architecture of the entrance guard system is designed according to the actual demand.The hardware platform of the embedded entrance is set up with the raspberry school development board as the core.The USB camera is used as the image acquisition device,and the peripheral control circuit is designed to cooperate with the potentiometer to complete the opening and closing of the electronic lock.On the basis of the introduction of tensorflow deep learning framework and opencv visual library,the FACENET algorithm based face recognition server,Web management platform,face detection program based on MTCNN algorithm,remote network request program,and door lock control program are developed.It implements functions such as door lock control,face registration,access control,attendance and so on.At the same time,administrators can access remote management platform anytime and anywhere.3.In the actual application scenario test the function and performance of the system,and analyze the factors affecting the performance.The experimental results show that the system can meet the expected standard of the actual application scene in the face detection rate,the face recognition rate and the time consuming,and the rich function can meet the daily needs.The system design has excellent performance and good application prospects.