Using Deep Learning Model To Identify Iron Chlorosis In Plants

This project introduces a Vision Transformer (ViT)-based deep learning approach for detecting and classifying iron chlorosis in plant leaves. Iron chlorosis, a major nutrient deficiency, leads to leaf yellowing, reducing crop yield and quality. Early and precise detection is crucial for effective intervention and precision agriculture. Unlike traditional models that focus on local image features, ViT captures long-range dependencies, enhancing feature extraction and classification. A curated dataset with four classes (Healthy, Mild, Moderate, and Severe Chlorosis) undergoes preprocessing, including resizing, normalization, and augmentation, to improve robustness. The model is fine-tuned using transfer learning with pre-trained weights, ensuring high accuracy in chlorosis detection. Evaluation on a test set demonstrates ViT's superiority over conventional methods. This automated system enables farmers and agricultural experts to assess plant health efficiently, offering real-time recommendations for managing iron chlorosis. The results confirm that ViT is a promising tool for precision agriculture and automated plant health monitoring.