Answers are grounded in this paper and cite their sources.
Plant diseases continue to be one of the leading causes of reduced agricultural productivity worldwide, directly threatening food supply chains and the economic stability of farming communities. With the global population steadily increasing, the demand for intelligent, scalable, and highly accurate plant disease detection systems has never been more critical. Deep learning methods have shown promising results in this field; however, numerous conventional models cannot often generalize well across different crop species and unseen disease types. These limitations hinder their practical deployment in dynamic real-world agricultural environments. In this study, we propose a robust and generalized deep learning-based approach for cross-crop plant disease detection, using the comprehensive and diverse Plant Village dataset. Our model is built upon a custom-designed Convolutional Neural Network (CNN) architecture that incorporates a small Inception module. Unlike traditional CNNs, which primarily focus on the global features of a leaf. Our model detects and analyzes localized disease spread patterns, enhancing detection across diverse crops and adapting to novel conditions. The small Inception module plays a vital role in enabling multi-scale feature extraction from small disease-affected patches without adding excessive computational complexity. This architectural refinement allows the model to learn more discriminative features, resulting in faster convergence and higher classification accuracy. When trained and validated on the Plant Village dataset, our model achieved an impressive accuracy of 98.45%, outperforming many traditional approaches. Additionally, it demonstrated consistently high precision, recall, and F1-score, confirming its reliability and robustness. By addressing the challenges of overfitting and poor generalization, common pitfalls in many deep learning models, our method provides a scalable and effective solution for real-time agricultural disease monitoring. This work contributes to the growing field of precision agriculture by offering a model that is not only accurate but also generally efficient and practical for deployment in diverse agricultural settings. Ultimately, our research aims to support the development of smart farming technologies that ensure healthier crops and contribute to long-term global food security.
No takes yet. Share an insight, caveat, or question.
Roopa et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: