Improving the Performance of Detection and Classification of Maize Plant Diseases Using Deep Neural Networks on Large Image Datasets

Maize plants have grown popularly in several tropical countries. The plant has provided us with a high value in food and industry. However, some diseases have caused a great damage to the productivity of the plant. The paper presents a method to detect and recognized maize diseases using data augmentation and deep neural networks. The detection of diseased regions is performed using the YOLOv11. After that, the classification of maize diseases is performed using the EfficientNet V2 network. The proposed method for the detection and classification is evaluated on a large public dataset of maize plant diseases. Obtained detection accuracy is 93% and the classification accuracy is 96%. Obtained results have demonstrated the promising applications of our proposed method.