Tree Areas in Satellite Imagery
Description
This dataset is designed for binary classification tasks on geospatial imagery, specifically to distinguish between land areas with trees and those without. The images were captured by the Sentinel-2 satellite. The dataset structure is straightforward: - Each image has a resolution of 64×64 pixels with encoded in JPG format. - Images are organized into two folders: "Trees" and "NoTrees", corresponding to the two classes. - Each folder contains 5,200 images, totaling 10,400 images across the dataset. Note: The dataset does not include predefined training, validation, or test splits. Users should partition the data as needed for their specific machine learning, deep learning workflows.
Files
Steps to reproduce
- Unzip the trees-in-satellite-imagery.zip - Discover the folders named "Trees" and "NoTrees" folder.
Institutions
- Marmara Universitesi