Akademik Veri Yönetim Sistemi
Physiological and Molecular Plant Pathology, cilt.144, 2026 (SCI-Expanded, Scopus)
Potato late blight, caused by Phytophthora infestans , remains one of the most destructive plant diseases worldwide, demanding rapid detection, accurate forecasting, and timely intervention. Recent advances in artificial intelligence have enabled a new generation of plant-protection systems that integrate multimodal sensing, autonomous robotics, and real-time digital twin frameworks. This review provides a structured and critical synthesis of major artificial intelligence (AI) approaches applied to potato late blight management, including physiologically informed leaf-level detection, multimodal sensing, spatio-temporal forecasting, and emerging foundation and generative models. Advanced forecasting approaches, including LSTM networks, temporal CNNs, spatial transformers, and graph neural networks, are examined for their ability to capture the complex spatio-temporal dynamics of disease spread under fluctuating microclimates. Where available, comparative insights are highlighted across model classes, emphasizing performance trade-offs, data requirements, and deployment constraints under real-field conditions. Generative AI GANs, diffusion models, and synthetic weather simulators address critical data gaps by reproducing physiologically relevant lesion progression and rare environmental disease scenarios. Digital twins unify these components, creating dynamic virtual replicas of potato fields that assimilate sensor streams, simulate intervention strategies, and provide interpretable decision-support dashboards. The review also discusses robotics and UAV platforms for large-scale surveillance, edge-AI deployment, and autonomous variable-rate spraying. Finally, it highlights key challenges in data quality, model drift, ethics, and accessibility, while outlining a roadmap toward next-generation, region-wide intelligent plant-health ecosystems. Through this integrated perspective, the review demonstrates how AI is transforming late blight management from reactive control to predictive, simulation-driven crop protection.