The Design Of Household Near Infrared Face Recognition Access Control System Based On The Chip AM437X

With the rapid development of smart home system,intelligent door lock,as an important product in the field of smart home,has broad prospects for development.At present,intelligent door lock mostly used passwords,ID cards,and fingerprints to recognize identity,these identity authentication methods cannot meet the requirements of customers.As a noncontact biometric technology,face recognition is suitable for the intelligent door lock industry.Most mature face recognition access control systems on the market are designed for office buildings,airports,stations and other public places,it can not be directly applied to intelligent door locks.According to the characteristics of intelligent door lock,this paper designs a household near infrared face recognition access control system.This paper first defines the system functional requirements and performance indicators,and gives the overall design scheme.In order to achieve the balance between power consumption and performance,the system adopts a "dual-core" architecture.AM437 X is selected as the s ystem processor to complete the functions of face recognition,user interface,etc.STM32 L is selected as the system controller,which is responsible for managing the power supply of each module and controlling the underlying peripherals.The system uses near infrared face recognition,by combining near infrared camera with active light source,the quality of input image is improved,the illumination robustness is enhanced.In addition,this paper uses proximity induction circuit instead of traditional touc h switch,so that the system can automatically wake up from the dormant mode,complete the workflow such as face recognition,access control,wireless communication and so on,and the system does not need to be started by manual keys,which fully embodies the non-contact advantages of face recognition.Finally,the system was installed in the intelligent door lock mechanical structure and tested.The main experimental results are as follows: the start-up time is about 3 seconds;the face recognition time is less than 160 milliseconds;the sleep power consumption is less than 3 mW;and the average operating power consumption is less than 1.5 W.When the system is powered by a lithium battery with the capacity of 37 Wh,it can work for about 7500 hours with an average of 10 face recognition unlock operations per day.The results show that the system can achieve all the functional requirements,all the performance indicators have reached the design goals,and the system runs steadily.