Dataset of groundnut plant leaf images for classification and detection

Abstract

The use of machine learning is rapidly expanding across many industries, including agriculture and the IT sector. However, data is essential for machine learning models, and a substantial amount of data is required prior to training a model. We have collected data of groundnut plant leaves in the form of digital photographs taken in the Koppal (Karnataka, India) area with the assistance of a pathologist in natural settings. Images of leaves are categorized into six distinct groups according to their condition. Collected images are pre-processed and the processed images of groundnut leaves are kept in 6 folders as: the "healthy leaves" folder with 1871 images, the "early leaf spot" folder with 1731 images, the "late leaf spot" folder with 1896 images, the "Nutrition deficiency" folder with 1665 images, the "rust" folder with 1724 images, and the "early rust" folder with 1474 images. The total number of images in the dataset is 10361. This dataset will be useful to train and validate deep learning and machine learning algorithms for groundnut leaf disease classification and recognition. Disease detection in plants is crucial for limiting crop losses and our dataset will help disease detection in groundnut plants. This dataset is freely accessible to public at https://data.mendeley.com/datasets/22p2vcbxfk/3 and at https://doi.org/10.17632/22p2vcbxfk.3.

Plant phenotyping relevance

落花生葉の画像データセット自体を構築・公開し、葉の健康状態や病徴分類に利用する研究であり、植物病害状態の画像ベース表現型取得が中心です。

Code and data availability

The paper is a data descriptor for a public groundnut leaf image dataset (10,361 annotated images across six disease/health classes) deposited on Mendeley Data, with explicit public URLs and DOI. This is a paper-specific plant image dataset directly used for the phenotyping/disease-classification analysis. No author's