Investigating the Effectiveness of Deep Learning Models in Detecting Wheat Diseases

Timely and precise identification of wheat diseases is essential to guarantee food security and optimize agricultural output. This study examines the efficacy of different deep learning models in identifying and categorizing wheat illnesses based on leaf photos. We evaluate the accuracy, computational efficiency, and robustness of Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and hybrid designs. The dataset comprises annotated photos of wheat leaves afflicted with prevalent diseases such as rust, powdery mildew, and leaf blotch. Models undergo training and validation using conventional criteria such as precision, recall, F1-score, and total accuracy. In addition, we investigate how data augmentation and transfer learning approaches affect the performance of the model. The results of our study suggest that CNN-based models, especially those that incorporate transfer learning, demonstrate higher levels of accuracy and resilience in comparison to RNN and hybrid models. This study offers useful insights into the appropriateness of deep learning models for practical agricultural uses and emphasizes the potential of these technologies to aid in automated disease monitoring and management.