YOLO-Based Image Detection System for Early Detection of Tomato Plant Diseases

Abstract

Food security and sustainable agriculture rely heavily on the timely detection and classification of plant diseases. In this study, we investigate the performance of the You Only Look Once (YOLO) object detection algorithm—specifically versions v5, v7, and v8—for identifying seven common tomato leaf diseases: Mosaic Virus, Leaf Miner, Septoria, Spider Mites, Early Blight, Yellow Leaf Curl Virus, and Late Blight. We trained and validated each YOLO variant using a comprehensive dataset comprising annotated images of diseased tomato leaves. YOLOv8 achieved the highest performance, with a mean Average Precision (mAP) of 85%, followed by YOLOv7 (84.7%) and YOLOv5 (83%). Additionally, YOLOv8 demonstrated the fastest inference time, indicating its suitability for real-time or near real-time disease detection applications. Our findings emphasize YOLOv8’s potential in enhancing agricultural productivity through accurate and efficient disease identification. The proposed framework offers practical implications for precision farming by aiding early disease management, optimizing crop yield, conserving resources, and promoting sustainable agricultural practices through advanced deep learning techniques.

Plant phenotyping relevance

トマト葉の病徴を画像から検出・分類するYOLO手法を訓練・検証し、性能比較も行っており、植物病害状態の表現型取得が中心である。

Code and data availability

The preprint describes a 700-image tomato leaf disease dataset and YOLOv5/v7/v8 experiments, but provides no public deposit, repository, or availability statement for the dataset, annotations, code, or trained models. No paper-specific public asset is actionable, and no allowed URLs are provided.

No evidence-backed public reproduction asset is currently recorded.