metadata

task_categories:
  - image-classification

So2Sat

So2Sat is a local climate zone (LCZ) classification task consisting of images from Sentinel-1 and Sentinel-2. We only keep Sentinel-2 images which have 10 MSI bands in this dataset. In addition, we reserve 10% of the training set as the validation set, resulting in 31,713 training samples, 3,523 validation samples and 48,307 test samples.

How to Use This Dataset

from datasets import load_dataset

dataset = load_dataset("GFM-Bench/So2Sat")

Also, please see our GFM-Bench repository for more information about how to use the dataset! 🤗

Dataset Metadata

The following metadata provides details about the Sentinel-2 imagery used in the dataset:

Number of Sentinel-2 Bands: 10
Sentinel-2 Bands: B02 (Blue), B03 (Green), B04 (Red), B05 (Vegetation red edge), B06 (Vegetation red edge), B07 (Vegetation red edge), B08 (NIR), B8A (Narrow NIR), B11 (SWIR), B12 (SWIR)
Image Resolution: 32 x 32 pixels
Spatial Resolution: 10 meters
Number of Classes: 17

Dataset Splits

The So2Sat dataset consists following splits:

train: 31,713 samples
val: 3,523 samples
test: 48,307 samples

Dataset Features:

The So2Sat dataset consists of following features:

optical: the Sentinel-2 image.
label: the classification label.
optical_channel_wv: the central wavelength of each Sentinel-2 bands.
spatial_resolution: the spatial resolution of images.

Citation

If you use the So2Sat dataset in your work, please cite the original paper:

@article{zhu2019so2sat,
  title={So2Sat LCZ42: A benchmark dataset for global local climate zones classification},
  author={Zhu, Xiao Xiang and Hu, Jingliang and Qiu, Chunping and Shi, Yilei and Kang, Jian and Mou, Lichao and Bagheri, Hossein and H{\"a}berle, Matthias and Hua, Yuansheng and Huang, Rong and others},
  journal={arXiv preprint arXiv:1912.12171},
  year={2019}
}