Automated Soybean Crop Health Evaluation from UAV Images Using Patch Level CNNs

Abstract - The rapid advancement of unmanned aerial vehicles (UAVs) and deep learning techniques has significantly transformed crop monitoring and precision agriculture. Among various crops, soybean plays a crucial role in global food and oilseed production, making timely and accurate crop health assessment essential. This review presents a comprehensive analysis of UAV-based soybean crop health evaluation methods, with a particular focus on image-based disease detection and stress monitoring using deep learning models. Recent progress in convolutional neural networks, patch-level image analysis, attention mechanisms, and lightweight architectures is systematically examined. The paper discusses commonly used UAV imaging modalities, preprocessing strategies, model architectures, and evaluation practices reported in the literature. Furthermore, existing challenges such as environmental variability, computational complexity, data imbalance, and real-world deployment constraints are critically analyzed. Based on the reviewed studies, potential research directions are identified, emphasizing efficient patch-level learning, interpretable health mapping, and scalable field-level assessment. This review aims to provide researchers and practitioners with a clear understanding of current trends, limitations, and future opportunities in UAV-assisted soybean crop health monitoring. Key Words: - Unmanned Aerial Vehicles (UAVs), Soybean Crop Health Monitoring, Precision Agriculture, Deep Learning