<![CDATA[This study aims to develop a sorting algorithm for the SmartEdu Conveyor using computer vision technology to enhance accuracy and efficiency in automated sorting systems. The system integrates a Raspberry Pi 4 as the main processing unit and employs the YOLOv8 object detection algorithm to classify geometric objects moving on a conveyor belt. Images captured by an overhead camera are processed in real time, and the results are transmitted through the MQTT protocol using the Paho MQTT library. Node-RED functions as the Human-Machine Interface (HMI), while a Programmable Logic Controller (PLC) drives double-acting pneumatic cylinders to perform the sorting mechanism. Experimental tests conducted at three conveyor speeds demonstrate that the system achieves an average accuracy confidence of 89.38% at 1 cm/s, 78.57% at 1.7 cm/s, and 59.28% at 2.3 cm/s. Further performance evaluation using the Precision–Recall curve yields a mean Average Precision (mAP) of 0.993 at an Intersection over Union (IoU) threshold of 0.5, indicating highly accurate object detection capability. The proposed YOLOv8-based sorting system demonstrates reliable real-time operation, high precision, and robust communication between vision and control modules. It will be implemented as a SmartEdu teaching aid prototype to support automation learning and industrial training applications. This work contributes to educational automation by integrating an open-source vision algorithm with industrial control architecture.]]>
