NeuralCrop: Intelligent Plant Disease Detection with CNNs

Abstract

Plant diseases significantly threaten global agricultural productivity, necessitating rapid and accurate detection methods for effective crop yield management. Traditional identification approaches are often labor-intensive and require specialized knowledge. In this study, we leverage advanced deep learning techniques, specifically Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs), to enhance plant disease detection accuracy. Utilizing a meticulously collected multispectral dataset with six 50 mm filters, spanning both visible and near-infrared (NIR) wavelengths, we explore innovative methodologies for disease classification. achieving an overall accuracy of 90% with similar models. This comparative analysis underscores the critical impact of balanced datasets and optimal wavelength selection on the efficacy of deep learning models for robust disease identification. These findings not only promise to advance crop disease management practices in agricultural settings but also contribute to enhancing global food security. Our study emphasizes the transformative potential of machine learning in plant disease diagnostics and advocates for ongoing research in this vital area.