Advanced Deep Learning and Transformer-Based Approaches for Apple Leaf Disease Detection and Smart Agriculture Applications: A Comprehensive Review

Abstract

Apple leaf diseases greatly impact agricultural yield, crop quality, and the economic stability of current farming systems. Getting the disease spotted early, and accurately, is crucial for reducing crop losses and also for improving precision agriculture routines. Usual identification approaches still depend heavily on manual checking by specialists, and that is pretty slow, costly, and sometimes misread by humans , even when the effort is careful. In the last few years AI, ML, Deep Learning, plus computer vision tools, have shifted plant disease monitoring toward automated and more intelligent recognition pipelines. This review paper tries to cover advanced deep learning and Transformer centered methods for apple leaf disease detection, while paying special attention to YOLO families, Convolutional Neural Networks, ResNet based models, Swin Transformers, and hybrid deep learning frameworks. The discussion examines new progress related to YOLOv8 YOLOv10 YOLOv11, attention mechanisms, multi scale feature extraction, Bayesian optimization , and lightweight detectors that are made for real time field use. Also the review touches image preprocessing approaches, feature extraction techniques, data enlargement strategies, and hyperparameter tuning methods that collectively raise classification precision and detection quality. When models are compared, hybrid CNN Transformer designs tend to deliver better disease localization, stronger feature encoding , and higher classification accuracy in difficult weather and background conditions. The paper additionally points to smart farming integration, like IoT setups, UAV monitoring, and edge computing, to enable more automated agricultural surveillance. Toward the end, the current research gaps, practical limits, ongoing obstacles, and future work paths are laid out, with the goal of supporting efficient scalable, and intelligent plant disease detection systems, aimed at sustainable agriculture and precision farming.