Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -5,15 +5,17 @@ task_categories:
|
|
5 |
language:
|
6 |
- en
|
7 |
tags:
|
8 |
-
|
9 |
-
|
10 |
-
|
11 |
-
|
12 |
-
|
13 |
-
|
|
|
|
|
14 |
pretty_name: cloudsen12
|
15 |
size_categories:
|
16 |
-
|
17 |
---
|
18 |
|
19 |
|
@@ -38,32 +40,81 @@ those who prefer traditional storage formats, GeoTIFF files are available in our
|
|
38 |
|
39 |
*CloudSEN12+ spatial coverage. The terms p509 and p2000 denote the patch size 509 × 509 and 2000 × 2000, respectively. ‘high’, ‘scribble’, and ‘nolabel’ refer to the types of expert-labeled annotations*
|
40 |
|
41 |
-
|
|
|
|
|
|
|
|
|
42 |
|
43 |
-
|
44 |
|
|
|
45 |
|
46 |
-
|
47 |
-
|------|------|--------|---------------------------------|--------------------------|
|
48 |
-
| 0 | B1 | 0.0001 | 443.9 nm (S2A)/442.3 nm (S2B) | Aerosols. |
|
49 |
-
| 1 | B2 | 0.0001 | 496.6 nm (S2A)/492.1 nm (S2B) | Blue. |
|
50 |
-
| 2 | B3 | 0.0001 | 560 nm (S2A)/559 nm (S2B) | Green. |
|
51 |
-
| 3 | B4 | 0.0001 | 664.5 nm (S2A)/665 nm (S2B) | Red. |
|
52 |
-
| 4 | B5 | 0.0001 | 703.9 nm (S2A)/703.8 nm (S2B) | Red Edge 1. |
|
53 |
-
| 5 | B6 | 0.0001 | 740.2 nm (S2A)/739.1 nm (S2B) | Red Edge 2. |
|
54 |
-
| 6 | B7 | 0.0001 | 782.5 nm (S2A)/779.7 nm (S2B) | Red Edge 3. |
|
55 |
-
| 7 | B8 | 0.0001 | 835.1 nm (S2A)/833 nm (S2B) | NIR. |
|
56 |
-
| 8 | B8A | 0.0001 | 864.8 nm (S2A)/864 nm (S2B) | Red Edge 4. |
|
57 |
-
| 9 | B9 | 0.0001 | 945 nm (S2A)/943.2 nm (S2B) | Water vapor. |
|
58 |
-
| 10 | B10 | 0.0001 | 1373.5 nm (S2A)/1376.9 nm (S2B) | Cirrus. |
|
59 |
-
| 11 | B11 | 0.0001 | 1613.7 nm (S2A)/1610.4 nm (S2B) | SWIR 1. |
|
60 |
-
| 12 | B12 | 0.0001 | 2202.4 nm (S2A)/2185.7 nm (S2B) | SWIR 2. |
|
61 |
-
| 13 | CM1 | 1 | - | Expert-labeled image. |
|
62 |
-
| 14 | CM2 | 1 | - | UnetMobV2-labeled image. |
|
63 |
|
|
|
64 |
|
65 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
66 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
67 |
|
68 |
The **fixed/** folder contains high and scribble labels, which have been improved in this new version. These changes have already been integrated.
|
69 |
|
@@ -73,7 +124,48 @@ The **images/** folder contains the CloudSEN12+ imagery
|
|
73 |
|
74 |
## Download
|
75 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
76 |
|
77 |
## Citation
|
78 |
|
79 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
language:
|
6 |
- en
|
7 |
tags:
|
8 |
+
- clouds
|
9 |
+
- earth-observation
|
10 |
+
- remote-sensing
|
11 |
+
- sentinel-2
|
12 |
+
- deep-learning
|
13 |
+
- multi-spectral
|
14 |
+
- satellite
|
15 |
+
- geospatial
|
16 |
pretty_name: cloudsen12
|
17 |
size_categories:
|
18 |
+
- 100K<n<1M
|
19 |
---
|
20 |
|
21 |
|
|
|
40 |
|
41 |
*CloudSEN12+ spatial coverage. The terms p509 and p2000 denote the patch size 509 × 509 and 2000 × 2000, respectively. ‘high’, ‘scribble’, and ‘nolabel’ refer to the types of expert-labeled annotations*
|
42 |
|
43 |
+
**ML-STAC Snippet**
|
44 |
+
```python
|
45 |
+
import mlstac
|
46 |
+
dataset = mlstac.load('isp-uv-es/CloudSEN12Plus')
|
47 |
+
```
|
48 |
|
49 |
+
**Sensor: Sentinel2 - MSI**
|
50 |
|
51 |
+
**ML-STAC Task: image-segmentation**
|
52 |
|
53 |
+
**Data raw repository: [https://cloudsen12.github.io/](https://cloudsen12.github.io/)**
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
54 |
|
55 |
+
**Dataset discussion: [https://huggingface.co/datasets/isp-uv-es/CloudSEN12Plus/discussions](https://huggingface.co/datasets/isp-uv-es/CloudSEN12Plus/discussions)**
|
56 |
|
57 |
+
**Split_strategy: stratified**
|
58 |
+
|
59 |
+
**Paper: [https://www.sciencedirect.com/science/article/pii/S2352340924008163](https://www.sciencedirect.com/science/article/pii/S2352340924008163)**
|
60 |
+
|
61 |
+
|
62 |
+
## Data Providers
|
63 |
+
|
64 |
+
|Name|Role|URL|
|
65 |
+
| :---: | :---: | :---: |
|
66 |
+
|Image & Signal Processing|['host']|https://isp.uv.es/|
|
67 |
+
|ESA|['producer']|https://www.esa.int/|
|
68 |
+
|
69 |
+
## Curators
|
70 |
+
|
71 |
+
|Name|Organization|URL|
|
72 |
+
| :---: | :---: | :---: |
|
73 |
+
|Cesar Aybar|Image & Signal Processing|http://csaybar.github.io/|
|
74 |
+
|
75 |
+
|
76 |
+
## Labels
|
77 |
+
|
78 |
+
|Name|Value|
|
79 |
+
| :---: | :---: |
|
80 |
+
|clear|0|
|
81 |
+
|thick-cloud|1|
|
82 |
+
|thin-cloud|2|
|
83 |
+
|cloud-shadow|3|
|
84 |
+
|
85 |
+
## Dimensions
|
86 |
+
|
87 |
+
|Axis|Name|Description|
|
88 |
+
| :---: | :---: | :---: |
|
89 |
+
|0|C|Spectral bands|
|
90 |
+
|1|H|Height|
|
91 |
+
|2|W|Width|
|
92 |
|
93 |
+
## Spectral Bands
|
94 |
+
|
95 |
+
|Name|Common Name|Description|Center Wavelength|Full Width Half Max|Index|
|
96 |
+
| :---: | :---: | :---: | :---: | :---: | :---: |
|
97 |
+
|B01|coastal aerosol|Band 1 - Coastal aerosol - 60m|443.5|17.0|0|
|
98 |
+
|B02|blue|Band 2 - Blue - 10m|496.5|53.0|1|
|
99 |
+
|B03|green|Band 3 - Green - 10m|560.0|34.0|2|
|
100 |
+
|B04|red|Band 4 - Red - 10m|664.5|29.0|3|
|
101 |
+
|B05|red edge 1|Band 5 - Vegetation red edge 1 - 20m|704.5|13.0|4|
|
102 |
+
|B06|red edge 2|Band 6 - Vegetation red edge 2 - 20m|740.5|13.0|5|
|
103 |
+
|B07|red edge 3|Band 7 - Vegetation red edge 3 - 20m|783.0|18.0|6|
|
104 |
+
|B08|NIR|Band 8 - Near infrared - 10m|840.0|114.0|7|
|
105 |
+
|B8A|red edge 4|Band 8A - Vegetation red edge 4 - 20m|864.5|19.0|8|
|
106 |
+
|B09|water vapor|Band 9 - Water vapor - 60m|945.0|18.0|9|
|
107 |
+
|B10|cirrus|Band 10 - Cirrus - 60m|1375.5|31.0|10|
|
108 |
+
|B11|SWIR 1|Band 11 - Shortwave infrared 1 - 20m|1613.5|89.0|11|
|
109 |
+
|B12|SWIR 2|Band 12 - Shortwave infrared 2 - 20m|2199.5|173.0|12|
|
110 |
+
|CM1| Cloud Mask 1| Expert-labeled image. |-|-|13|
|
111 |
+
|CM2| Cloud Mask 2| UnetMobV2-V1 labeled image. |-|-|14|
|
112 |
+
|
113 |
+
## Data Structure
|
114 |
+
|
115 |
+
We use `.mls` format to store the data in HugginFace and GeoTIFF for ScienceDataBank.
|
116 |
+
|
117 |
+
## Folder Structure
|
118 |
|
119 |
The **fixed/** folder contains high and scribble labels, which have been improved in this new version. These changes have already been integrated.
|
120 |
|
|
|
124 |
|
125 |
## Download
|
126 |
|
127 |
+
The code below can be used to download the dataset using the `mlstac` library. For a more detailed example, please refer to the `examples` section in our
|
128 |
+
website [https://cloudsen12.github.io/](https://cloudsen12.github.io/).
|
129 |
+
|
130 |
+
|
131 |
+
```python
|
132 |
+
import mlstac
|
133 |
+
import matplotlib.pyplot as plt
|
134 |
+
import numpy as np
|
135 |
+
|
136 |
+
ds = mlstac.load(snippet="isp-uv-es/CloudSEN12Plus")
|
137 |
+
subset = ds.metadata[(ds.metadata["split"] == "test") & (ds.metadata["label_type"] == "high") & (ds.metadata["proj_shape"] == 509)][10:14]
|
138 |
+
datacube = mlstac.get_data(dataset=subset)
|
139 |
+
```
|
140 |
+
|
141 |
+
Make a plot of the data point downloaded
|
142 |
+
|
143 |
+
```python
|
144 |
+
datapoint = datacube[2]
|
145 |
+
datapoint_rgb = np.moveaxis(datapoint[[3, 2, 1]], 0, -1) / 5_000
|
146 |
+
fig, ax = plt.subplots(1, 3, figsize=(10, 5))
|
147 |
+
ax[0].imshow(datapoint_rgb)
|
148 |
+
ax[0].set_title("RGB")
|
149 |
+
ax[1].imshow(datapoint[13], cmap="gray")
|
150 |
+
ax[1].set_title("Human label")
|
151 |
+
ax[2].imshow(datapoint[14], cmap="gray")
|
152 |
+
ax[2].set_title("UnetMobV2 v1.0")
|
153 |
+
```
|
154 |
+
|
155 |
+
|
156 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6402474cfa1acad600659e92/scVhZf3rkB3uWkZZ6Epmu.png)
|
157 |
|
158 |
## Citation
|
159 |
|
160 |
+
Cite the dataset as:
|
161 |
+
|
162 |
+
```bibtex
|
163 |
+
@article{aybar2024cloudsen12+,
|
164 |
+
title={CloudSEN12+: The largest dataset of expert-labeled pixels for cloud and cloud shadow detection in Sentinel-2},
|
165 |
+
author={Aybar, Cesar and Bautista, Lesly and Montero, David and Contreras, Julio and Ayala, Daryl and Prudencio, Fernando and Loja, Jhomira and Ysuhuaylas, Luis and Herrera, Fernando and Gonzales, Karen and others},
|
166 |
+
journal={Data in Brief},
|
167 |
+
pages={110852},
|
168 |
+
year={2024},
|
169 |
+
publisher={Elsevier}
|
170 |
+
}
|
171 |
+
```
|