Deep Learning-Based Plant Disease Detection from Leaf Images

Our study aims to discover a new method for detecting plant diseases using Deep Learning (DL) algorithms. The data was also collected from the global Kaggle website, and contains more than 25,000 images. A number of important plants in daily life were tested, such as tomato, corn and potatoes. There were many common diseases in these plants, which lead to lack of production and damage to agricultural crops. Among them, for the tomato plant, there was leaf spot disease, mold, bacterial spots, and many diseases. As for the corn plant, there was brown and root spot disease and bacteria. For potatoes, there were several diseases. These include early blight, common rust, northern leaf blight, and late blight, and each of these diseases negatively affects the crop and poses a threat to plant life. In this study, we used a number of MobileNet and ResNet50 algorithms to detect plant diseases, as these algorithms demonstrated a high ability to detect plant diseases.