Identificação de doenças em folhas de plantas de soja utilizando uma rede neural convolucional

Abstract

Soybean, as one of the most important crops in the world, faces several challenges related to diseases that compromise its productivity and quality. This study aims to explore the application of a machine learning technique, based on a convolutional neural network, for identifying specific diseases in soybean plants through leaf images. The model was developed using a dataset of images labeled by experts, representing different leaf conditions, including diseases such as Asian rust, target spot, and powdery mildew, as well as healthy leaves. The dataset was divided into training and testing sets. Performance was evaluated using metrics such as accuracy, which achieved a result of 70%, F1-Score, with a result of 0.89, and the confusion matrix. The results, particularly the accuracy, were lower than those found in related studies. However, the datasets used in those studies are larger, which is believed to be the main reason for the lower value in this work.