Sistema de controle e gerenciamento de acesso

Abstract

Technological advances have allowed for the development of increasingly advanced physical access control systems, including electronic devices such as password systems, RFID, and biometrics. Among the most popular biometric systems are facial recognition and fingerprinting, which compare an individual’s biometric characteristics with data previously stored in a database. This paper proposes the creation of an access control and management system that integrates different technologies, including electronic passwords, RFID tags, and facial recognition. In addition, the development of a web-based data management system is proposed. The work was developed using the Raspberry Pi 2 model b, MFRC552 RFID module, 4X4 matrix keypad, camera, button, MariaDB for database management, programming languages Python, HTML, CSS, and PHP, and technologies such as Google Colab. The developed modules function independently and communicate through a socket API. The facial recognition module was validated using Google Colab and was developed using the Long Short Term Memory (LSTM) neural network, achieving an accuracy of 97.04%. After the processes were created, it was necessary to integrate the technologies. Unfortunately, due to hardware limitations, it was not possible to implement facial recognition verification. However, the necessary machine learning algorithms and models for facial recognition were developed and tested on a computer using the MOBIO database. The final system includes password and RFID tag entry options, and despite this limitation, the obtained results were satisfactory, allowing access control only for registered individuals.