<![CDATA[Smart LED display systems are widely used to provide useful information to users, ranging from simple LED screens to complex screens management and monitoring systems involving a large number of diverse devices, capable of integrating modern technologies. This research focuses on developing a smart LED display management and monitoring system for a laboratory using AI-IoT technologies, which combines deep learning, computer vision, edge computing, embedded system, IoT Communication (MQTT), and web-based management. The goal is to provide convenience, efficiency, and flexibility for users and managers, enabling easy remote information updates and real-time display on LED screens, while simultaneously automatically monitoring and accurately counting the number of people entering and leaving the laboratory. The development of the system includes designing an ESP32-based central LED control board, selecting the P2.5 LED modules, the jetson nano, the Logitech C505e camera, suitable for low-cost educational research. Subsequently, the article introduces the image processing algorithm for counting people based on YOLOv7 TensorRT inference and develops the web management interface based on the Next.js platform, combined with data communication via MQTT protocol. This research was then experimentally implemented at the Mitsubishi FA Laboratory at the university of transport and communications (UTC). The experimental results showed that the Web interface features a grid layout divided into three functional groups, allowing for display content configuration, graphical visualization, clear status display. It provides networked link-tags for updating date/time, temperature/humidity, and In/Out people counts in real-time on both the Web and the LED screen via MQTT/ WebSocket protocols. The experimental results also indicated that the proposed algorithm for counting people In/Out the laboratory achieves high accuracy, over 90%, under normal, stable lighting conditions. This confirms that the proposed smart LED display system operates efficiently, stably, and reliably, and suitable for promoting the digital management of laboratories at a low investment cost.]]>
