Revolutionizing Grape Quality Control: Harnessing Artificial Intelligence for Enhanced Precision

In this study, a convolutional neural network (CNN) architecture adapted from the Tiny VGG model is proposed for automated disease detection in grapevine leaves, thus addressing challenges in vineyard phytosanitary management by grape-producing regions of Peru, such as Ica and Piura. Data augmentation techniques were used and three experiments were carried out to evaluate the performance of the model in classifying three fungal diseases on grapevine leaves. Additionally, the PlantVillage dataset was used for training. The results show that, with the basic configuration, the model achieved a higher accuracy (89.9%) and weighted F1 score (89.9%), highlighting the viability of this approach in the early and effective detection of diseases in viticulture.