A Cutlery Sorting Robot with AI Imaging Technology and Designed End-Effector

This paper presents a fully automated cutlery sorting system that integrates image processing and mechanical design to efficiently classify and grasp stacked metal cutlery. The system employs the YOLOv8 segmentation model for real-time object detection, accurately identifying the type of cutlery. To address challenges posed by unordered stacking and reflective surfaces, we design a tactile height detection mechanism combining an electromagnet and a micro switch to ensure precise cutlery catching. The principal component analysis (PCA) method is used to identify the orientation of each cutlery so that all cutlery can be grasped with a suitable posture and neatly arranged in the collecting box. Experimental results show that our approach achieves a 96% success rate across the entire process, from realtime detection to sorting, providing a reliable and efficient solution for automated cutlery classification in real-world applications.