metadata

license:
  - unknown
task_categories:
  - image-classification
language:
  - en
tags:
  - remote-sensing
  - earth-observation
  - geospatial
  - satellite-imagery
  - scene-classification
pretty_name: RSSCN7
size_categories:
  - 1K<n<10K

RSSCN7

Description

The RSSCN7 dataset is designed for scene classification tasks and provides a collection of high-resolution RGB images. This dataset comprises a total of 2,800 images, each with a resolution of 400x400 pixels. The images are extracted from Google Earth and cover a range of diverse scenes. RSSCN7 includes seven distinct scene classes, with 400 images per class.

Details

Statistics

Total Number of Images: 2,800
Image Resolution: 400x400 pixels
Scene Classes: 7
Dataset Size: 0.36GB

Citation

If you use the RSSCN7 dataset in your research, please consider citing the following publication or the dataset's official website:

@article{7272047,
  title        = {Deep Learning Based Feature Selection for Remote Sensing Scene Classification},
  author       = {Zou, Qin and Ni, Lihao and Zhang, Tong and Wang, Qian},
  year         = 2015,
  journal      = {IEEE Geoscience and Remote Sensing Letters},
  volume       = 12,
  number       = 11,
  pages        = {2321--2325},
  doi          = {10.1109/LGRS.2015.2475299}
}

Paper with code: https://paperswithcode.com/dataset/rsscn7
Repo: https://github.com/palewithout/RSSCN7