Computer Vision-Based Coffee Capsule Recognition and Robotic Manipulation for Automated Coffee Brewing

This paper presents an intelligent robotic system that automatically executes the coffee brewing process. The system contains a 6 -axis robotic arm, a camera, a laptop, and a capsule coffee machine. The robot has a dual-function gripper that can grasp a coffee capsule and hold a cup without changing the gripper. Accurate capsule identification is achieved using the YOLOv8 model, and the chosen capsule posture is recognized using a specific method. The whole process is as follows. The user selects a desired coffee on his/her cell phone, and the robot receives the order and identifies the chosen coffee capsule. Then, the robotic arm moves to grasp the desired capsule according to its posture, bring it to the coffee machine, put it into the coffee machine, and press the brew button. Once the coffee flows into the cup in front of the machine, the system awaits the next command. The whole process will be completed automatically. The whole process technique includes capsule type identification and posture recognition, coordinate transformation, and robotic arm control with a user-friendly web-based interface. The experiment demonstrates that the system can be run well and accurately. This robot brewing system indeed offers a practical implementation for household automation.