Deep Learning-Based Wheat Disease Detection and Classification System Using Convolutional Neural Networks

Abstract

Wheat, one of the major crops in the world, is vulnerable to many diseases that cause tremendous yield and quality loss. This paper proposes a deep learning method for the automatic detection and classification of wheat diseases based on a Convolutional Neural Network (CNN). We respond to the imperative of early and precise identification of diseases in wheat crops in order to reduce agricultural losses.The system learned on a data set of more than 14,000 wheat leaf images corresponding to 15 classes of various rusts, blights, insects, and normal leaves. Our suggested CNN model reached a training accuracy of 97.02% and validation accuracy of 91.00%. The model design uses data augmentation strategies and dropout regularization to promote generalization as well as avoid overfitting