Smartphone image dataset to distinguish healthy and unhealthy leaves in papaya orchards in Bangladesh.

Abstract

Papaya, renowned for its nutritional benefits, represents a highly profitable crop. However, it is susceptible to various diseases that can significantly impede fruit productivity and quality. Among these, leaf diseases pose a substantial threat, severely impacting the growth of papaya plants. Consequently, papaya farmers frequently encounter numerous challenges and financial setbacks. To facilitate the easy and efficient identification of papaya leaf diseases, a comprehensive dataset has been assembled. This dataset, comprising approximately 1400 images of diseased, infected, and healthy leaves, aims to enhance the understanding of how these ailments affect papaya plants. The images, meticulously collected from diverse regions and under varying weather conditions, offer detailed insights into the disease patterns specific to papaya leaves. Stringent measures have been taken to ensure the dataset's quality and enhance its utility. The images, captured from multiple angles and boasting high resolution are designed to aid in the development of a highly accurate model. Additionally, RGB mode has been employed to meticulously capture each detail, ensuring a flawless representation of the leaves. The dataset meticulously identifies and categorizes five primary types of leaf diseases: Leaf Curl (inclusive of its initial stage), Papaya Mosaic, Ring Spot, Mites (specifically, those affected by Red Spider Mites), and Mealybug. These diseases are recognized for their detrimental effects on both the leaves and the overall fruit production of the papaya plant. By leveraging this curated dataset, it is possible to train a model for the real-time detection of leaf diseases, significantly aiding in the timely identification of such conditions.

Plant phenotyping relevance

パパイヤ葉の健全・病害状態を画像で取得したデータセットであり、植物病害表現型の再利用可能なデータ基盤として中心的です。

Code and data availability

The paper is a Data in Brief article describing a smartphone image dataset of healthy and diseased papaya leaves (~1400 original, 6618 augmented images across six classes) collected in Bangladesh. The authors' own dataset is publicly deposited on Mendeley Data with an explicit direct URL and DOI, making it a paper-phen