Dataset for Detecting Pumpkin Leaf Diseases

Published: 18 August 2025| Version 1 | DOI: 10.17632/4xcm6y3dns.1
Contributors:
,
,

Description

The dataset contains a total of 11,936 images representing five classes of Pumpkin (Cucurbita) leaves: Powdery Mildew, Downy Mildew, Mosaic Virus, Bacterial Leaf Spot, and Healthy leaves. Images were captured in real farm environments in Gobindapur, Faridganj, Chandpur, Bangladesh, under varied lighting conditions including natural sunlight, artificial light, and mobile flash. The distribution of leaf samples is as follows: Healthy : Original - 329 samples Augmented - 2303 samples Powdery Mildew : Original - 257 samples Augmented - 1799 samples Downy Mildew : Original - 294 samples Augmented - 2058 samples Mosaic Virus : Original - 304 samples Augmented - 2128 samples Bacterial Leaf Spot : Original - 308 samples Augmented - 2156 samples

Files

Steps to reproduce

We Travel over half a month to collect data. Images were collected from Gobindapur, Faridganj, Chandpur, Bangladesh. Diseased and healthy leaves were identified under expert guidance. Photos were taken using Xiaomi 11T (108 MP) and Samsung A34 (48 MP) smartphone cameras. Only mature leaves showing clear symptoms or healthy appearance were included.

Institutions

  • Daffodil International University

Categories

Computer Vision, Precision Agriculture, Plant Diseases, Deep Learning

Licence