Hybrid Deep Learning Model for Date Palm Disease Classification: A Fusion of HybridConv Mixer and Vision Transformer

Area of Science:

• Agricultural Science
• Computer Vision
• Plant Pathology

Background:

• Date palm cultivation is vital for arid regions, but diseases like brown spots and white scale significantly reduce yield.
• Manual disease detection is labor-intensive, prone to errors, and inefficient for large-scale agricultural operations.

Purpose of the Study:

• To develop an automated system for accurate classification of date palm leaf diseases.
• To enhance the reliability and efficiency of disease detection in date palm cultivation.

Main Methods:

• A novel ensemble model combining Hybrid Convolutional Mixer (HybridConv) and Vision Transformer (ViT) was developed.
• The model was trained on a dataset of healthy and diseased date palm leaf images, including data augmentation for improved robustness.
• The HybridConv component focused on local feature detection, while ViT handled global feature analysis.

Main Results:

• The developed HybridConv-ViT model achieved a high accuracy of 99.89% in classifying brown spots, white scale, and healthy date palm leaves.
• The ensemble model outperformed traditional single Convolutional Neural Networks (CNN) models in precision, recall, and F1-score.
• Data augmentation further increased the model's reliability in disease identification.

Conclusions:

• The HybridConv-ViT ensemble model offers a highly accurate and efficient automated solution for date palm disease detection.
• This technology has the potential to significantly improve disease management and productivity in date palm agriculture.
• Automated disease classification systems are crucial for sustainable agricultural practices in challenging environments.