Enhanced Tomato Leaf Disease Detection using YOLOv11: A Deep Learning-based Approach

Early detection of plant diseases is very crucial for Indian agriculture, where the majority of the country's population is dependent. Many tomato crops in India are prone to leaf diseases, which greatly reduces crop yield if timely action is not taken. Therefore, we developed a system using YOLOv11n, a deep learning model, and trained it with images from the PlantVillage dataset, which has pictures of healthy tomato leaves along with nine different leaf diseases. In order to see how the YOLOv11n model affects tomato leaf disease diagnosis, we evaluated its results with and without using data augmentation. We made several changes to the dataset images during the training phase with the help of the Albumentations library to add geometric and colour changes to the images, which made the proposed model more reliable, robust, and accurate. It allows the model to give reliable predictions even though the images are taken under poor lighting, background, and different angles. The proposed model is able to detect the disease with Mean Average Precision at 0.5 (mAP@0.5) of 99.49%, recall of 99.49%, (mAP@0.5:0.95) of 99.48% and precision of 99.77%. A simple web application is developed such that farmers can easily upload leaf images and get the results of their crop disease, which helps farmers to manage their crops more effectively.