AI-Driven Mustard Disease Identification: A Multiclass and Binary Classification Approach for Advanced Crop Health Monitoring

Important for both food and oil production, mustard plants are frequently susceptible to maladies that can result in significant yield losses. Early detection and diagnosis of these diseases are essential for effective management and control, thereby contributing to the maintenance of the production chain and crop yield. In this study, we present a deep learning-based method for detecting mustard disease using leaf images. We have trained and evaluated a convolutional neural network (CNN) model dedicated to classifying mustard leaves as healthy or diseased. This article focuses on the binary classification of healthy and infected mustard crop samples. In addition, we have conducted multi-classification concentrating on the classification of five main mustard diseases, including stem rot, white rust, bacterial blight, powdery mildew, and downy mildew. This article is the first of its kind to perform classification concerning this particular mustard disease strain. The model provides adequate recall, precision, and f1 score performance. In precision agriculture, the proposed method can be used to automatically detect disease and manage crops to increase yield and crop quality as the predicted precision for healthy and unhealthy disease classification is 95.17% and 93.8% respectively.