Automated Vision-Based System for Comprehensive and Real-Time Detection and Classification of Fruit Surface Damages

Global agriculture faces increasing pressure to reduce food waste and enhance productivity to meet market demands. This project investigates the use of YOLOv8 for detecting fruit damage, providing a more efficient alternative to labor-intensive manual inspections. By utilizing high-resolution images of both healthy and damaged fruits, the YOLO-based deep neural network is trained to identify and localize fruit defects with high accuracy. Through transfer learning, the model is fine-tuned for fruit-specific damage detection, enabling rapid and reliable quality control. This automated approach reduces food waste, boosts efficiency, and ensures compliance with global market standards, minimizing financial risks and enhancing exporters' market competitiveness. The real-time capabilities of YOLOv8 make it particularly suited for large-scale agricultural operations, addressing key challenges in the sector while strengthening global trade.