Optimizing Apple Leaf Disease Detection with Advanced Ensemble Architectures

Apple leaf diseases pose a significant threat to the productivity and health of apple crops. This study explores advanced ensemble deep learning models, including VGG16 + ResNet50, VGG19 + InceptionV4, and ResNet50 + InceptionV4, for robust classification of apple leaf diseases. Using the Plant Pathology 2020 Challenge Dataset, the proposed models achieved outstanding performance, with the ResN et50 + Inception V 4 ensemble attaining an accuracy of 99.00%, precision of 0.99, and Kappa score of 0.986. The results highlight the potential of ensemble techniques in improving classification accuracy, addressing generalizability challenges, and enabling scalable solutions for agricultural diagnostics. Future directions include optimizing computational efficiency for real-time applications.