Deep Learning–Based Soybean Leaf Disease Classification : A Comprehensive Review

Abstract

Soybean is one of the most economically important crops worldwide, yet its productivity is severely affected by a wide range of foliar diseases. Traditional disease diagnosis relies on expert visual inspection, which is time-consuming, subjective, and impractical for large-scale monitoring. In recent years, computer vision and artificial intelligence have emerged as promising tools for automated soybean leaf disease classification. This review presents a comprehensive analysis of state-of-the-art image-based soybean leaf disease classification techniques, with particular emphasis on deep learning, transfer learning, federated learning, and transformer-based models. The paper systematically examines preprocessing strategies, feature extraction methods, classification architectures, and evaluation protocols used in recent studies. Furthermore, comparative insights are drawn across convolutional neural networks, lightweight models, ensemble approaches, and vision transformers. The review also highlights key research findings, existing challenges, and unresolved limitations such as dataset imbalance, real-field variability, explainability, and deployment constraints. By synthesizing recent advances and identifying open research directions, this review aims to guide researchers toward more robust, scalable, and intelligent soybean disease diagnosis systems for sustainable agriculture.