Detection Plant Diseases Using Deep Learning Algorithms

This study aimed to propose a detection approach for plant disease based on deep learning (DL) algorithms. The study sought to discover diseases affecting three plants, which are: Common Rust, Vercospora Leaf Spot, Northern Leaf Blight for Corn Plant, And Early Blight, Late Blight for Potato, And Bacterial Spot, Septoria Leaf Spot, Target Spot for Tomato. To implement this approach, three deep learning algorithms were used including VGG16, VGG19, and CNN. This model was trained on a dataset contain 25272 image from the Kaggle database. The results demonstrated that the proposed approach could accurately detect plant diseases. In comparison, the VGG19 algorithm has a high detection accuracy of up to 95%, while the VGG16 algorithm has a detection accuracy of up to 86%. Then followed the CNN algorithm, which had detection accuracy up to as 86%.