This project focuses on developing an innovative system for monitoring visitor flow in historical monuments, aiming to preserve their structural and cultural integrity. The primary subject of the study is the Convento de Cristo, a renowned monument located in Tomar, Portugal. With increasing tourist numbers, traditional manual methods of visitor counting and management have proven inadequate. The project seeks to replace these methods with a real-time automated system capable of accurately counting and tracking visitors. To achieve this goal, advanced computer vision algorithms were integrated, namely YOLOv5 for object detection and DeepSORT for real-time object tracking. The system architecture was designed for modularity and scalability, utilizing a Raspberry Pi 5 for video capture and Docker to containerize the machine learning stack. A user-friendly interface was developed using Flask, allowing users to monitor real-time visitor counts, visualize historical data, and manage system configurations with ease. Throughout the project, extensive testing was conducted using both pre-recorded video samples and live camera feeds to evaluate the system’s performance. Sockets were employed to enable efficient communication between the Raspberry Pi and the machine learning stack, ensuring real-time data processing. The system demonstrated the capability to accurately track individuals and adapt to various monitoring scenarios, with an average processing speed of approximately 40 frames per second and a delay of under one second. These results validate the proposed solution’s effectiveness for real-time monitoring.