Classification Of Wheat Disease Using CNN

Abstract

Leaf rust, stem rust, and powdery mildew are just a few of plant diseases that may drastically reduce the productivity of wheat, a cereal crop that is crucial on a worldwide scale.  In order to intervene quickly and implement sustainable farming techniques, early and precise disease diagnosis is essential.  Automated disease classification in wheat utilizing CNN is presented in this paper as a deep learning-based technique.  In order to make the model more generalizable, it was preprocessed and enhanced using a dataset of wheat leaf photos that included images of various diseases. Input photos were used to train a bespoke CNN architecture that could learn hierarchical features. Model's strong performance in differentiating amongst infected and healthy wheat samples was shown by its high classification accuracy.  In terms of accuracy, recall, and total F1-score, CNN model outperforms conventional machine learning classifiers.  With suggested approach, agronomists and farmers may have access to a scalable and effective tool for early disease identification in wheat, which would allow for more precise treatment and less yield loss.