Advancing Digital Agriculture: Deep Learning Based A Smart Rose Disease Classifier

Abstract

A Deep Learning-Based Smart Rose Disease Classifier tries to solve the problem of plant illnesses that are getting worse and hurting farming, costing farmers much money. This project is about using deep learning to quickly and accurately diagnose diseases that affect roses so that they can be treated early and less money is lost. Handling rare diseases by hand requires much knowledge, so the process is complex, takes a long time, and needs much work. This project aims to use image processing better to make a dataset from a nearby rose garden. The collection has 24,801 pictures, grouped into seven groups: Rust, Fresh Leaf, Botrytis blight, Cercospora leaf spot, Downy mildew, and Black spot. Three deep learning models—AlexNet, CNN, and MobileNet V2—were used for training and testing. The MobileNet V2 algorithm did the best, with a fantastic accuracy rate of 98.86%