Intelligent Fruit and Vegetable Counting System Based on Raspberry Pi 4B and SG-YOLOv8
摘要
With the growth of unmanned and intelligent retail, traditional manual weighing and pricing have become inefficient and costly. This paper presents an automated fruit and vegetable pricing system based on an improved YOLOv8 model. The SG-YOLOv8 integrates a StarNet backbone, GhostConv modules, and the lightweight LSCD detection head to reduce computation while maintaining accuracy. Deployed on Raspberry Pi 4B with ESP32, Tkinter, and MySQL, the system achieves 95.9% detection accuracy, boosts efficiency by 40%, and cuts costs by 70%, demonstrating strong practicality and effectiveness.