Design And Implementation Of Intelligent Access Control System Based On Edge Computing

With the rise and development of the Internet of Things,intelligent access control based on face recognition is gradually being applied by more and more companies.At present,most intelligent access control systems can only identify internal employees provided by customers,and cannot respond flexibly to external visitors who need temporary enter.In addition,many access controls follow mainstream Internet of Things technologies,separating "computing" and "operations",requiring higher transmission channels and storage capacity,and also affecting the response speed.In order to improve the above problems,the author's company set up the Smart Edge platform project team to develop smart access control applications for external visitors.Edge computing utilizes the computing and processing capabilities of the edge device itself to complete some of the computing tasks that originally belonged to the computing center.This not only improves the response speed of edge applications,but also reduces latency,improves user experience,and eases communication pressure,ensuring data security.The system is based on the idea of edge computing and combined with machine learning algorithms.It separates the face recognition process of external visitors from the model training process before recognition,and is deployed in edge measurement and cloud measurement,respectively,to achieve cloud-based collaborative common services.In addition,with the important communication protocol MQTT of the Internet of Things as the core,intelligent edge tools are deployed on the edge devices,and the edge devices are managed through the communication between the management platform and the edge tools.The author participated in the project needs analysis during the project development process to clarify the business needs of the project.After completing the requirements analysis,the author divides the project functional modules and outlines the system according to the functional modules.After that,the author designed the architecture of specific modules,the logic flow between modules and the database tables,and used Flask network framework,Docker containerization technology,ssm network framework,MQTT protocol,MYSQL database,python container management interface and other technologies to complete the development and implementation of the visitor registration module,model training module,end-side identification module,device scheduling platform and intelligent edge tools.Finally,the deployment of the project is completed.At present,the system has started trial in the author's company.External visitors can submit a visit application through the We Chat mini program.After passing the application,they can directly brush their faces and enter the company at the company's entrance.The administrator can also manage the access control device through the management platform,and update the application version and model data in the device in real time.