A Review on Plant Leaf Disease Detection Using CNN Deep Learning Models

Abstract

This paper aims to review the advancements in early-stage plant disease detection using neural network-based approaches, with a particular focus on improving agricultural practices and disease management. The objective is to provide a comprehensive overview of the various neural network architectures applied in plant disease detection, their methodologies, and the results achieved in real-world applications. The review encompasses the use of deep learning models, particularly convolutional neural networks (CNNs), for classifying and detecting plant diseases from images. It also discusses the preprocessing techniques, feature extraction methods, and dataset variations employed in these systems. Various evaluation metrics such as accuracy, precision, recall, and F1-score are analyzed to compare the performance of different models. The results presented in the review highlight the potential of neural networks in providing accurate, automated disease detection solutions that can assist farmers in early intervention. These systems contribute to enhancing crop yield, reducing pesticide use, and promoting sustainable farming practices. However, challenges such as dataset diversity, the need for extensive training data, and real-time application remain. The paper concludes by emphasizing the importance of developing robust, scalable, and easy-to-deploy models for real-time disease detection, with the potential to revolutionize precision agriculture and improve food security across diverse farming environments. Future research directions include enhancing model generalization, integrating multi-modal data, and improving the interpretability of deep learning models for practical agricultural use.