An Innovative Deep Learning-based Diagnostic Approach for Apple Leaf Disease Detection Assisted by Ensemble Learning

Apple is one of the world's staple foods, consumed by over a billion people. However, a number of diseases that affect the quality and quantity of apples are posing a growing threat to apple production. These diseases, which are frequently caused by climate, soil type, pests, and pathogens, have a major impact on crop health, and among them, apple leaf diseases—like Apple Scab, Powdery Mildew, and Cedar Apple Rust—are particularly destructive, resulting in a decline in apple tree health and overall productivity. To overcome these challenges, scientists are investigating cutting-edge technologies such as deep learning for image processing to detect and classify diseases of apple leaves. This makes it possible for farmers to make timely adjustments, which increases yields and productivity. The proposed study offers a comprehensive analysis of previous research and suggests a model for the automatic identification of apple leaf diseases with the use of ensemble learning and transfer learning techniques. With a testing accuracy of 99.52%, the created methodology outperforms existing advanced models in terms of accuracy, precision, recall, and F1-score, including VGG19 (99.32%), ResNet50 (98.32%), InceptionV3 (98.32%), MobileNet (97.96%), and AlexNet (97.63%). The outcomes show the suggested model's advantage in crucial evaluation criteria in addition to its high accuracy.