A Comprehensive Mango Leaf Disease Dataset with Semantic Annotations

Published: 29 January 2025| Version 1 | DOI: 10.17632/278k27zd4t.1
Contributors:
Jahin Morshed,

Description

The Mango Leaf Disease dataset comprises a collection of images of mango leaves gathered from an orchard in Ramgarh, Khagrachari, Bangladesh, specifically from Sierra Agro, located at the geographical coordinates 22.98123°N and 91.76185°E. The data collection took place between October 10 and October 12, 2024. The dataset contains a total of 2,731 images, including both healthy leaves and diseased leaves. These images are categorized into five distinct classes: Anthracnose, Dag Disease, Red Rust, Galls, and Healthy. The photographs were taken using the high-resolution camera of an iPhone 12 under natural lighting conditions on sunny days, with temperatures ranging from 28°C to 30°C. The original images were captured in HEIF format and later converted to JPG format, with dimensions of 2543x3390 pixels. For practical use, the images have been resized to 640x480 pixels with a resolution of 72 dpi. Additionally, the dataset includes detailed annotations that categorize the diseases, facilitating semantic segmentation, a feature that has garnered significant attention from data scientists.

Files

Institutions

  • Daffodil International University

Categories

Agricultural Science, Computer Science, Artificial Intelligence, Computer Vision, Machine Learning, Pattern Recognition, Deep Learning

Licence