AI-Driven Identification of Grapevine Fungal Spores via Microscopic Imaging and Feature Optimization with Cuckoo Search Algorithm
Grapevine diseases caused by fungal pathogens pose a significant threat to viticulture, leading to considerable economic losses and reduced productivity. Early, intelligent detection of fungal spores is vital for effective disease management. This study presents a high-accuracy classification model...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-08-01
|
| Series: | Smart Agricultural Technology |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S277237552500262X |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Grapevine diseases caused by fungal pathogens pose a significant threat to viticulture, leading to considerable economic losses and reduced productivity. Early, intelligent detection of fungal spores is vital for effective disease management. This study presents a high-accuracy classification model that uses microscopic images to differentiate among four closely related grapevine pathogens: Lasiodiplodia brasiliensis, L. crassispora, L. exigua, and L. gilanensis. Advanced image processing techniques were applied to segment spores and extract texture, color, and shape features. A support vector machine (SVM) classifier achieved 97.5% overall accuracy after preprocessing, a substantial improvement over the initial 81.25% without image preprocessing. Individual species classification accuracies were 95.24% for L. brasiliensis, 95.00% for L. crassispora, and 100% for both L. exigua and L. gilanensis. Feature selection using the cuckoo search algorithm identified ten key attributes for classification. Among these, texture features (energy, contrast, entropy, homogeneity, and standard deviation) proved most impactful, achieving classification accuracies of 75.50%, 70.06%, 62.14%, 55.67%, and 52.51%, respectively. Shape features (major and minor axis lengths, area) and color features (mean, standard deviation) followed with lower accuracy scores. The results emphasize the supportive role of texture in fungal spore classification and highlight the supporting roles of shape and color. This AI-driven framework offers a robust tool for early disease diagnosis in grapevines, with broader implications for automated plant pathology, precision agriculture, and smart farming technologies. |
|---|---|
| ISSN: | 2772-3755 |