Deep Learning Techniques for Microscopic Identification and Detection of Plant-Parasitic Nematodes
JOURNAL OF CROP HEALTH, cilt.78, sa.5, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Derleme
- Cilt numarası: 78 Sayı: 5
- Basım Tarihi: 2026
- Doi Numarası: 10.1007/s10343-026-01409-8
- Dergi Adı: JOURNAL OF CROP HEALTH
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
- Eskişehir Osmangazi Üniversitesi Adresli: Evet
Özet
Accurate identification of plant-parasitic nematodes (PPNs) and determination of their population densities are of great importance for reducing yield losses in agricultural production. However, traditional morphological and molecular diagnostic methods of PPNs are both time-consuming and require expertise. Recent advances in deep learning (DL) have created new opportunities to automate PPN detection more rapidly and reliably. This review evaluates DL approaches applied to the classification, detection, and counting of PPNs, with emphasis on convolutional neural network (CNN) architectures, transfer-learning frameworks, and modern object-detection models such as YOLO, which are notable for their real-time detection capacity. Studies in the literature reveal that DL models are highly successful in distinguishing PPN species from microscopic images and can rapidly identify their life stages. These methods contribute to making the diagnostic process more stable, repeatable, and accessible by reducing nematologist-dependent errors. Nevertheless, limitations in dataset size, variation in imaging conditions across laboratories, and morphological differences associated with developmental stages of PPNs make it difficult for models to adapt to all conditions. Therefore, the establishment of larger and better-annotated datasets, as well as the integration of different imaging techniques with DL models, emerges as an important necessity for the future. In conclusion, DL-based approaches offer significant advantages in terms of speed, accuracy, and scalability for PPN detection; when supported by an appropriate data infrastructure, they become a powerful tool for sustainable plant health management.