Wheat Rust Disease Detection and Classification using an improved Deep Learning Algorithm

Abstract Wheat, the third most widely consumed cereal crop worldwide, faces substantial yield and quality losses as a result of rust disease, notably leaf rust, stem rust, and stripe rust. These rust disease, caused by Puccinia triticina , Puccinia graminis , and Puccinia striiformis , respectively, are capable of causing significant yield losses in wheat in the absence of timely detection. Conventional disease identification relies heavily on manual visual inspection, which is time consuming, labor intensive, and prone to error, especially in large scale agricultural systems. To address these limitations, this study proposes a deep learning-based framework for the early detection and classification of wheat rust diseases. A real-time dataset was developed using field images collected from various wheat-growing regions and augmented with publicly available data. The dataset comprises images of healthy leaves and those affected with the three major rust diseases. A modified convolutional neural network (CNN) architecture was employed for extract features and disease classification. Experimental results demonstrate that the proposed approach achieves high classification accuracy, highlighting its effectiveness as a reliable tool for automated wheat rust detection in precision agriculture. By enabling rapid and accurate disease identification, the system supports timely decision-making, reduces potential yield losses, and improves crop management practices, thereby contributing to food security and sustainable agricultural production.