TOM2024: Datasets of tomato, onion, and maize images for developing pests and diseases AI-based classification models.

Abstract

The advancement of digital technologies has significantly impacted plant pest and disease management, yet gaps remain, especially in developing regions. This paper introduces the TOM2024 dataset, a comprehensive collection of high-resolution images designed to enhance pest and disease identification of maize, tomato, and onion crops. The dataset encompasses 25,844 raw images and over 12,000 labeled images, categorized into 30 classes (healthy crop, infested crop, and pest) across the three cropping systems. Acquired through meticulous fieldwork in Burkina Faso using high-resolution cameras, the dataset includes diverse environmental conditions and crop stages, ensuring a robust resource for AI model training and validation. The dataset is segmented into three categories: processed images (Category A), selected images with augmentation (Category B), and an online repository with over 25,000 raw images (Category C). Category A and B features images of crops affected by 21 distinct pests and diseases. This dataset addresses critical gaps in existing collections by offering extensive coverage and high-resolution imagery that can be used to developed AI models for automatic identification and classification of pests and diseases that affects crops. TOM2024's versatility extends to research, educational purposes, and the practical application of digital tools in agriculture thereby contributes to the advancement of precision agriculture, sustainable agricultural practices, and food security globally.

Plant phenotyping relevance

植物の健全・感染状態を含む画像データセットを構築し、病害・害虫状態の自動分類モデル開発用リソースとして提供することが中心であり、再利用可能な画像ベース表現型データセットに該当する。

Code and data availability

The paper is a Data in Brief article describing the TOM2024 dataset of tomato, onion, and maize pest/disease images, publicly deposited on Mendeley Data with an explicit direct URL and DOI. This is a paper-specific public image dataset (phenotyping-style plant image asset) directly produced by this paper.