Technological advances have allowed the development of increasingly advanced physical access control systems. Like a wide variety of electronic devices, the most recent systems stand out for their ability to recognize specific features. This paper presents the development and implementation of an access control and management system using computer vision techniques for license plate recognition. The main objective of the project is to teach High School students in Electronics at Federal Institute of Amazonas, Campus Manaus Distrito Industrial (IFAM-CMDI), Brazil, about practical and theoretical concepts related to the Internet of Things (IoT), with special emphasis on the use of computer vision for system autonomy. In this sense, we used Project-Based Learning (PBL) and Maker Culture, which are based on the principle that students learn more effectively when they are actively involved in the process, reflecting, discussing, collaborating, and applying knowledge in practical and real contexts, and interacting with a postgraduate student in IoT. This approach provided practical and integrated learning, allowing students to explore various modern technologies such as Jetson Nano and MariaDB for database management, as well as programming languages such as Python, HTML, CSS and PHP. In the context of computer vision, the project uses EasyOCR for character recognition and YOLOv8 for accurate identification of vehicle license plates. The system architecture includes a web server for access management and real-time identification and recognition of license plates. The results obtained were satisfactory, tested with a set of images obtained from the Internet, with different levels of quality, angle and distance, achieving an accuracy of 77%. The results obtained show that the interaction between students of different educational levels contributed to the mutual learning of students.

Keywords: Project-Based Learning, Computational Vision, Internet of Things.