Utilizing Machine Learning for Disease Detection in Mango Trees

This research aims to identify the various diseases and pests that afflict mango trees, with a particular focus on the leaves. The leaves serve as a clear indicator of pest impact, providing a vivid picture of the type of disease or pest affecting the tree. In this study, we have categorized eight distinct types of diseases and pests. To accurately identify these categories, this research conducted experiments using seven Convolutional Neural Network (CNN) models, enhanced with optimized hyperparameters. This approach is designed to achieve optimal results in disease recognition based on observable leaf conditions. Our experiments with the EfficientNet and VGG 16 models have been notably successful, achieving an impressive accuracy rate of over 90% in identifying the eight categories of diseases and pests. Additionally, the VGG 19 model demonstrated exceptional performance, reaching an accuracy rate of up to 99%. While this research has been largely successful in identifying diseases in mango trees, it is currently limited to these eight classes. Future research could expand to include other types of tree diseases, thereby broadening and deepening our understanding of this critical area.