Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
task_categories:
|
| 3 |
+
- image-segmentation
|
| 4 |
+
---
|
| 5 |
+
# MARIDA
|
| 6 |
+
**MARIDA** is a dataset for sparsely labeled marine debris which consists of 11 MSI bands. This dataset contains a training set of 694 samples along with a validation set of 328 samples and a test set of 350 samples. All image samples are originally 256 x 256 pixels. Wecombine both the original validation set and test set into one single test set (678 samples). Weemploy the same approach as DFC2020’s where we divide 256 x 256 pixels into 9 smaller patches of 96 x 96 pixels. Thus, our final training set contains 5,622 training samples, 624 validation samples and 6,183 test samples. All images are 96 x 96 pixels.
|
| 7 |
+
|
| 8 |
+
## How to Use This Dataset
|
| 9 |
+
```python
|
| 10 |
+
from datasets import load_dataset
|
| 11 |
+
|
| 12 |
+
dataset = load_dataset("GFM-Bench/MARIDA")
|
| 13 |
+
```
|
| 14 |
+
|
| 15 |
+
Also, please see our [GFM-Bench](https://github.com/uiuctml/GFM-Bench) repository for more information about how to use the dataset! 🤗
|
| 16 |
+
|
| 17 |
+
## Dataset Metadata
|
| 18 |
+
|
| 19 |
+
The following metadata provides details about the Sentinel-2 imagery used in the dataset:
|
| 20 |
+
<!--- **Number of Sentinel-1 Bands**: 2-->
|
| 21 |
+
<!--- **Sentinel-1 Bands**: VV, VH-->
|
| 22 |
+
- **Number of Sentinel-2 Bands**: 11
|
| 23 |
+
- **Sentinel-2 Bands**: B01 (**Coastal aerosol**), 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**)
|
| 24 |
+
- **Image Resolution**: 96 x 96 pixels
|
| 25 |
+
- **Spatial Resolution**: 10 meters
|
| 26 |
+
- **Number of Classes**: 11
|
| 27 |
+
|
| 28 |
+
## Dataset Splits
|
| 29 |
+
The **MARIDA** dataset consists following splits:
|
| 30 |
+
- **train**: 5,622 samples
|
| 31 |
+
- **val**: 624 samples
|
| 32 |
+
- **test**: 6,183 samples
|
| 33 |
+
|
| 34 |
+
## Dataset Features:
|
| 35 |
+
The **MARIDA** dataset consists of following features:
|
| 36 |
+
<!--- **radar**: the Sentinel-1 image.-->
|
| 37 |
+
- **optical**: the Sentinel-2 image.
|
| 38 |
+
- **label**: the segmentation labels.
|
| 39 |
+
<!--- **radar_channel_wv**: the central wavelength of each Sentinel-1 bands.-->
|
| 40 |
+
- **optical_channel_wv**: the central wavelength of each Sentinel-2 bands.
|
| 41 |
+
- **spatial_resolution**: the spatial resolution of images.
|
| 42 |
+
## Citation
|
| 43 |
+
If you use the MARIDA dataset in your work, please cite the original paper:
|
| 44 |
+
```
|
| 45 |
+
@article{kikaki2022marida,
|
| 46 |
+
title={MARIDA: A benchmark for Marine Debris detection from Sentinel-2 remote sensing data},
|
| 47 |
+
author={Kikaki, Katerina and Kakogeorgiou, Ioannis and Mikeli, Paraskevi and Raitsos, Dionysios E and Karantzalos, Konstantinos},
|
| 48 |
+
journal={PloS one},
|
| 49 |
+
volume={17},
|
| 50 |
+
number={1},
|
| 51 |
+
pages={e0262247},
|
| 52 |
+
year={2022},
|
| 53 |
+
publisher={Public Library of Science San Francisco, CA USA}
|
| 54 |
+
}
|
| 55 |
+
```
|