Deep Learning-Based Early Detection of Plant Diseases for Precision Agriculture

Abstract

Crop diseases threaten global agricultural production and food supply, creating a need for efficient early detection systems. Traditional detection methods are slow and labor-intensive, highlighting the need for automated solutions using machine learning. This study evaluates deep learning models, specifically Artificial Neural Networks (ANN) and Convolutional Neural Networks (CNN), for plant disease classification from leaf images using a high-resolution dataset with 38 disease classes. The CNN model, which preserves spatial information through convolutional layers, achieved 97.85% validation accuracy, while the ANN model, which treats images as one-dimensional vectors, showed lower performance. The structure, training, and evaluation of both models demonstrate the effectiveness of CNNs for automated crop disease diagnosis and early detection.