Freshness Classification Using YOLO

Abstract

This research explores a novel deep-learning approach for automating the classification of fruits and vegetables based on freshness using YOLOv8, a state-of-the-art object detection model by Ultralytics. Traditional manual inspection methods are time-consuming, subjective, and prone to human error. This paper proposes a YOLOv8-based system that addresses these limitations by offering real-time, objective detection and classification of produce freshness.

We leverage a curated dataset acquired from Kaggle, encompassing a diverse range of fruits and vegetables with varying freshness levels. Through rigorous training, the YOLOv8 model learns to identify visual cues associated with freshness, such as color, surface texture, and blemishes. The model's performance is evaluated on unseen data, demonstrating its ability to categorize accurately produced into different freshness classes.

The proposed system presents significant potential for various applications within the food industry. It can streamline quality control processes by automating freshness assessment, optimize inventory management by identifying produce nearing expiration, and empower consumers with real-time information about product freshness through mobile applications. Enhancing efficiency and accuracy in freshness classification contributes to improved food safety, reduced waste throughout the supply chain, and, ultimately, a more sustainable food system.