AI-Powered Strawberry Quality Assessment: A Breakthrough in Precision Agriculture

Strategic Insights -

Researchers at Paulista University in Brazil have made a significant contribution to the field of precision agriculture with a study on the use of artificial intelligence (AI) to assess the quality of strawberries. The study, published in the AgriEngineering journal, presents a deep learning-based approach for automated quality assessment using the YOLOv8n object detection model. The model was trained on a custom dataset of 5663 annotated strawberry images, covering eight quality categories, including anthracnose, gray mold, powdery mildew, uneven ripening, and physical defects.

Key Takeaways:

  • The study proposes a deep learning-based approach for automated quality assessment of strawberries using the YOLOv8n object detection model, achieving a mean Average Precision (mAP) of 0.79 and an inference time of 1 ms per image.
  • The model was trained on a custom dataset of 5663 annotated strawberry images, covering eight quality categories, including anthracnose, gray mold, powdery mildew, uneven ripening, and physical defects.
  • Data augmentation techniques, such as rotation and Gaussian blur, were applied to enhance model generalization and robustness.
  • The 200-epoch model achieved the best results, with a mAP50 of 0.79, demonstrating suitability for real-time applications.
  • Classes with distinct visual features, such as anthracnose and gray mold, were accurately classified, while visually similar categories, such as 'Good Quality' and 'Unripe' strawberries, presented classification challenges.
  • The study has significant implications for the precision agriculture industry, enabling farmers to make data-driven decisions to improve crop quality and reduce postharvest losses.
  • The research was conducted by Luana dos Santos Cordeiro, Irenilza de Alencar Naas, and Marcelo Tsuguio Okano, and was published in the AgriEngineering journal.

Statistics:

  • 5663 annotated strawberry images were used to train the YOLOv8n object detection model.
  • The model achieved a mean Average Precision (mAP) of 0.79 and an inference time of 1 ms per image.
  • The 200-epoch model achieved the best results, with a mAP50 of 0.79.
  • The study covers eight quality categories, including anthracnose, gray mold, powdery mildew, uneven ripening, and physical defects.

Sources:

  • AgriEngineering, 2025,7(8):246 (Smart Postharvest Management of Strawberries: YOLOv8-Driven Detection of Defects, Diseases, and Maturity)
  • Luana dos Santos Cordeiro, Irenilza de Alencar Naas, Marcelo Tsuguio Okano (Paulista University, Sao Paulo, Brazil)