---
language: en
license: unknown
task_categories:
- change-detection
pretty_name: ChaBuD MSI
tags:
- remote-sensing
- earth-observation
- geospatial
- satellite-imagery
- change-detection
- sentinel-2
dataset_info:
  features:
  - name: image1
    dtype:
      array3_d:
        dtype: uint8
        shape:
        - 512
        - 512
        - 13
  - name: image2
    dtype:
      array3_d:
        dtype: uint8
        shape:
        - 512
        - 512
        - 13
  - name: mask
    dtype: image
  splits:
  - name: train
    num_bytes: 2624716428.0
    num_examples: 278
  - name: validation
    num_bytes: 736431228.0
    num_examples: 78
  download_size: 2232652835
  dataset_size: 3361147656.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
---

# ChaBuD MSI

<!-- Dataset thumbnail -->
![ChaBuD MSI](./thumbnail.png)

<!-- Provide a quick summary of the dataset. -->
ChaBuD is a dataset for Change detection for Burned area Delineation and is used for the ChaBuD ECML-PKDD 2023 Discovery Challenge. This is the MSI version with 13 bands.
- **Paper:** https://doi.org/10.1016/j.rse.2021.112603
- **Homepage:** https://huggingface.co/spaces/competitions/ChaBuD-ECML-PKDD2023

## Description

<!-- Provide a longer summary of what this dataset is. -->


- **Total Number of Images**: 356
- **Bands**: 13 (MSI)
- **Image Size**: 512x512
- **Image Resolution**: 10m
- **Land Cover Classes**: 2
- **Classes**: no change, burned area
- **Source**: Sentinel-2


## Usage

To use this dataset, simply use `datasets.load_dataset("blanchon/ChaBuD_MSI")`.
<!-- Provide any additional information on how to use this dataset. -->
```python
from datasets import load_dataset
ChaBuD_MSI = load_dataset("blanchon/ChaBuD_MSI")
```

## Citation 

<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->
If you use the ChaBuD_MSI  dataset in your research, please consider citing the following publication:


```bibtex
@article{TURKOGLU2021112603,
    title = {Crop mapping from image time series: Deep learning with multi-scale label hierarchies},
    journal = {Remote Sensing of Environment},
    volume = {264},
    pages = {112603},
    year = {2021},
    issn = {0034-4257},
    doi = {https://doi.org/10.1016/j.rse.2021.112603},
    url = {https://www.sciencedirect.com/science/article/pii/S0034425721003230},
    author = {Mehmet Ozgur Turkoglu and Stefano D'Aronco and Gregor Perich and Frank Liebisch and Constantin Streit and Konrad Schindler and Jan Dirk Wegner},
    keywords = {Deep learning, Recurrent neural network (RNN), Convolutional RNN, Hierarchical classification, Multi-stage, Crop classification, Multi-temporal, Time series},
}
```