Hybrid CNN-RNN Modelling for Precisely Classifying the Seriousness of Mustard White Rust Disease at Five Levels

Mustard white rust disease is a severe threat to the cultivation of mustard plants, influencing quality and yield. This study developed a hybrid Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) model to predict Mustard white rust disease severity in five predefined levels. A dataset of about 12000 captured and chosen myself was carefully arranged; it contained the ratio for training, validation, and test sets as grouped below. The overall accuracy of 95.45%, shows considerable strength in identifying differences between stages of different diseases. Changes in performance metrics, accuracy, precision, recall, and F1-score refined understanding of the model's strong points. It indicated high degrees of precision but some small limitations concerning recall; most notably due to frequent incidence among samples lacking disease characteristics. Analyzing the confusion matrix showed that although one of the model's advantages was predicting true positives, it sometimes had problems identifying severe disease. The study points to the model's potential and offers ways for its future refinement, including making it more sensitive, especially about severe forms of disease expression. In summary, the hybrid CNN-RNN model exhibits potential for Mustard white rust disease severity assessment. With this methodology in place, agricultural productivity can be improved and foster targeted disease management strategies.