Deep Learning Model Based on ResNet9 for Early Blight and Late Blight Disease Detection in Potato (#566)

Authors

Quispe Layme, Mario Ronald

Ccoyuri Apaza, Jeremy Jose

Abstract

The present study addresses the need to detect potato diseases such as early blight, late blight. The objective is to evaluate the performance and accuracy of the Deep Learning technique using in particular the ResNet-9 model convolutional neural network architecture for the identification of diseases on potato leaves. A dataset from the PlantVillage repository on Kaggle was used, which includes 12,009 images of potato leaves classified into three categories: early blight, late blight, and healthy leaves. The model was trained on Google Colab, achieving an accuracy of 99.55% at the final stage. The research highlights the importance of early disease detection to reduce losses in agricultural production, especially in a context where the world population is increasing and the demand for food is growing. Traditional diagnostic techniques, such as visual inspection, are limited and often inaccurate, making the use of Deep Learning a promising solution. This approach not only improves the accuracy of disease identification, but also offers an affordable alternative for low-income farmers who lack the resources for expensive diagnostic techniques. The study concludes that the implementation of Deep Learning algorithms, such as ResNet-9, can be highly effective for disease detection in crops, contributing to more efficient and sustainable agriculture.