Cafeteria Self-Service Billing System Based on YOLOv11 and HSV Threshold Segmentation Algorithm

In this paper, a cafeteria self-service billing system is designed to address the problems of low recognition speed, weak anti-interference ability, and poor accuracy, etc. This design flexibly combines the YOLOv11 target detection algorithm and the HSV threshold segmentation algorithm, and the YOLOv11 algorithm improves the backbone and neck architecture, which greatly improves the feature extraction capability. A new convolutional mechanism C3k2 and C2PSA, as well as the depth separable convolution DWConv are introduced, which successfully realizes high-precision recognition of dish types as well as faster recognition speed and stronger robustness, which can improve the user experience and system efficiency. And through experiments, the average recognition accuracy and confidence of YOLO v11 are 2.8% and 2.4% higher than YOLOv10, and 3.8% and 4.4% higher than YOLOv8, respectively, and the HSV threshold segmentation algorithm can also identify the color of the dinner plate very well. In this paper, the HSV threshold segmentation algorithm is also tested in different environments, and the experimental results can also prove that the recognition accuracy is still above 80% even in the case of strong reflection. The hardware hardware part of the system to Raspberry Pi 4b as the motherboard, selected PN532 near-field communication module and thus the use of NFC for the interaction of the comprehensive test results show that the cafeteria self-service billing system is stable, can identify the type of food and the color of the dinner plate and calculate the total price, the accuracy is high, but also has a very good robustness, anti-interference ability, real-time is good, and has a certain degree of theoretical research and practical application value.