Design And Implementation Of Face Recognition Access Control System Based On Embedded Platform

With the increasing demand for security,confidentiality,and epidemic prevention,many units,enterprises,and communities have adopted strict access control.Traditional access control systems use keys or swiping cards,which have disadvantages such as the inability to confirm the identity of the holder,all person need key,and the holder can copy key illegally.Therefore,the current access control system based on biometric identification has achieved rapid development and application.Face recognition technology is one of the most widely used technologies in the field of biometrics.It has the features of fast,intuitive,cost-effective,non-contact,etc.Therefore,face recognition technology has broad application prospects in file management,security verification,personnel tracking,etc.,and is especially suitable for use as identity verification of access control systems.At the same time,with the continuous development of the field of machine learning,face recognition technology base on deep neural networks has also been greatly improved in accuracy and robustness,but the computational complexity has also increased,which cause the difficulty to meet the requirements of real-time performance to run face recognition programs on embedded devices with limited computing power.In response to the above problems,this thesis designs and implements a face recognition access control system that can balance accuracy and real-time,deploy on embedded devices base on deep neural networks.The specific work mainly includes the following aspects:(1)We have studied the development of related technologies such as face recognition and living body detection.Based on the analysis of the requirements of the access control system,we proposed a design scheme of an embedded face recognition access control system with face recognition detection,identity management,and log recording and other functions.(2)The face detection function is realized through the MTCNN network,and face recognition is completed based on the lightweight Face Net model.According to the difference between the imaging effects of infrared cameras and ordinary cameras under different attack methods,a method of living body detection is proposed.By comparing the gaps of different optical flow calculation methods,a live detection function for photo attacks is realized.(3)The management module of the system is designed in detail,and the functions of the entire management module are introduced in detail with the user interface prototype diagram and key code examples.The functional test of the access control system based on the embedded platform is carried out,and the detection speed,accuracy and reliability of the system have been tested completely.The test results show that the realization of the system basically meets the expected requirements.