Artificial intelligence-based pollen classification machine in apiculture: design, implementation and evaluation

Gerdan Koc D., KOÇ C., Ucak Koc A.

Journal of the Science of Food and Agriculture, 2025 (SCI-Expanded) identifier identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Basım Tarihi: 2025
  • Doi Numarası: 10.1002/jsfa.70238
  • Dergi Adı: Journal of the Science of Food and Agriculture
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Periodicals Index Online, Aerospace Database, Agricultural & Environmental Science Database, Analytical Abstracts, Aquatic Science & Fisheries Abstracts (ASFA), BIOSIS, CAB Abstracts, Chemical Abstracts Core, Communication Abstracts, Food Science & Technology Abstracts, INSPEC, Metadex, Pollution Abstracts, Veterinary Science Database, DIALNET, Civil Engineering Abstracts
  • Anahtar Kelimeler: artificial intelligence, convolutional neural networks, deep learning, food safety, pollen classification, YOLOv8
  • Ankara Üniversitesi Adresli: Evet

Özet

BACKGROUND: Bee pollen is a bioactive substance valued for its nutritional and health-promoting properties. However, consistent quality control and classification are hampered by variability in its chemical and biological composition, which is largely dependent on botanical origin. Artificial intelligence offers an opportunity to overcome these limitations through automated and standardized classification approaches. RESULTS: A deep learning-driven system was developed to classify pollen samples based on color properties, a key visual biomarker of floral origin. Convolutional neural networks including MobileNet, InceptionV3, Xception, NasNet Large, DenseNet201 and YOLOv8 were evaluated. DenseNet201 achieved the highest classification accuracy (98.5%), while YOLOv8, integrated for real-time performance, reached 91.4% accuracy with rapid processing speed. Laboratory-scale validation confirmed the reliability of the system in differentiating monofloral pollen types. CONCLUSION: The proposed AI-based classification system provides a robust solution for the standardization and traceability of pollen products. Its real-time classification capability offers beekeepers a practical tool for sustainable and hygienic pollen collection, with potential applications across the food, pharmaceutical, nutraceutical and cosmetic industries. © 2025 Society of Chemical Industry.