An Ensemble Deep Learning Approach for Reliable and Scalable Lemon Leaf Disease Classification

The study presents an ensemble deep learning approach for classifying lemon leaf diseases using a dataset of 1,354 images divided into 9 classes, including one for healthy leaves. By combining pretrained models InceptionV3 and MobileNetV2, the ensemble achieved an accuracy of 99.27%, with adversarial training enhancing robustness against noisy data. This approach is significant for practitioners as it demonstrates effective strategies for improving disease detection in agriculture using advanced model architectures and ensemble techniques.