Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,122 @@
|
|
| 1 |
---
|
| 2 |
license: cc-by-nc-4.0
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: cc-by-nc-4.0
|
| 3 |
---
|
| 4 |
+
|
| 5 |
+
# **CloudSEN12 - scribble**
|
| 6 |
+
## **A Benchmark Dataset for Cloud Semantic Understanding**
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
CloudSEN12 is a LARGE dataset (~1 TB) for cloud semantic understanding that consists of 49,400 image patches (IP) that are
|
| 11 |
+
evenly spread throughout all continents except Antarctica. Each IP covers 5090 x 5090 meters and contains data from Sentinel-2
|
| 12 |
+
levels 1C and 2A, hand-crafted annotations of thick and thin clouds and cloud shadows, Sentinel-1 Synthetic Aperture Radar (SAR),
|
| 13 |
+
digital elevation model, surface water occurrence, land cover classes, and cloud mask results from six cutting-edge
|
| 14 |
+
cloud detection algorithms.
|
| 15 |
+
|
| 16 |
+
CloudSEN12 is designed to support both weakly and self-/semi-supervised learning strategies by including three distinct forms of
|
| 17 |
+
hand-crafted labeling data: high-quality, scribble and no-annotation. For more details on how we created the dataset see our
|
| 18 |
+
paper.
|
| 19 |
+
|
| 20 |
+
Ready to start using **[CloudSEN12](https://cloudsen12.github.io/)**?
|
| 21 |
+
|
| 22 |
+
**[Download Dataset](https://cloudsen12.github.io/download.html)**
|
| 23 |
+
|
| 24 |
+
**[Paper - Scientific Data](https://www.nature.com/articles/s41597-022-01878-2)**
|
| 25 |
+
|
| 26 |
+
**[Inference on a new S2 image](https://colab.research.google.com/github/cloudsen12/examples/blob/master/example02.ipynb)**
|
| 27 |
+
|
| 28 |
+
**[Enter to cloudApp](https://github.com/cloudsen12/CloudApp)**
|
| 29 |
+
|
| 30 |
+
**[CloudSEN12 in Google Earth Engine](https://gee-community-catalog.org/projects/cloudsen12/)**
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
<br>
|
| 34 |
+
|
| 35 |
+
### **Description**
|
| 36 |
+
|
| 37 |
+
<br>
|
| 38 |
+
|
| 39 |
+
| File | Name | Scale | Wavelength | Description | Datatype |
|
| 40 |
+
|---------------|-----------------|--------|------------------------------|------------------------------------------------------------------------------------------------------|----------|
|
| 41 |
+
| L1C_ & L2A_ | B1 | 0.0001 | 443.9nm (S2A) / 442.3nm (S2B)| Aerosols. | np.int16 |
|
| 42 |
+
| | B2 | 0.0001 | 496.6nm (S2A) / 492.1nm (S2B)| Blue. | np.int16 |
|
| 43 |
+
| | B3 | 0.0001 | 560nm (S2A) / 559nm (S2B) | Green. | np.int16 |
|
| 44 |
+
| | B4 | 0.0001 | 664.5nm (S2A) / 665nm (S2B) | Red. | np.int16 |
|
| 45 |
+
| | B5 | 0.0001 | 703.9nm (S2A) / 703.8nm (S2B)| Red Edge 1. | np.int16 |
|
| 46 |
+
| | B6 | 0.0001 | 740.2nm (S2A) / 739.1nm (S2B)| Red Edge 2. | np.int16 |
|
| 47 |
+
| | B7 | 0.0001 | 782.5nm (S2A) / 779.7nm (S2B)| Red Edge 3. | np.int16 |
|
| 48 |
+
| | B8 | 0.0001 | 835.1nm (S2A) / 833nm (S2B) | NIR. | np.int16 |
|
| 49 |
+
| | B8A | 0.0001 | 864.8nm (S2A) / 864nm (S2B) | Red Edge 4. | np.int16 |
|
| 50 |
+
| | B9 | 0.0001 | 945nm (S2A) / 943.2nm (S2B) | Water vapor. | np.int16 |
|
| 51 |
+
| | B11 | 0.0001 | 1613.7nm (S2A) / 1610.4nm (S2B)| SWIR 1. | np.int16 |
|
| 52 |
+
| | B12 | 0.0001 | 2202.4nm (S2A) / 2185.7nm (S2B)| SWIR 2. | np.int16 |
|
| 53 |
+
| L1C_ | B10 | 0.0001 | 1373.5nm (S2A) / 1376.9nm (S2B)| Cirrus. | np.int16 |
|
| 54 |
+
| L2A_ | AOT | 0.001 | - | Aerosol Optical Thickness. | np.int16 |
|
| 55 |
+
| | WVP | 0.001 | - | Water Vapor Pressure. | np.int16 |
|
| 56 |
+
| | TCI_R | 1 | - | True Color Image, Red. | np.int16 |
|
| 57 |
+
| | TCI_G | 1 | - | True Color Image, Green. | np.int16 |
|
| 58 |
+
| | TCI_B | 1 | - | True Color Image, Blue. | np.int16 |
|
| 59 |
+
| S1_ | VV | 1 | 5.405GHz | Dual-band cross-polarization, vertical transmit/horizontal receive. |np.float32|
|
| 60 |
+
| | VH | 1 | 5.405GHz | Single co-polarization, vertical transmit/vertical receive. |np.float32|
|
| 61 |
+
| | angle | 1 | - | Incidence angle generated by interpolating the ‘incidenceAngle’ property. |np.float32|
|
| 62 |
+
| EXTRA_ | CDI | 0.0001 | - | Cloud Displacement Index. | np.int16 |
|
| 63 |
+
| | Shwdirection | 0.01 | - | Azimuth. Values range from 0°- 360°. | np.int16 |
|
| 64 |
+
| | elevation | 1 | - | Elevation in meters. Obtained from MERIT Hydro datasets. | np.int16 |
|
| 65 |
+
| | ocurrence | 1 | - | JRC Global Surface Water. The frequency with which water was present. | np.int16 |
|
| 66 |
+
| | LC100 | 1 | - | Copernicus land cover product. CGLS-LC100 Collection 3. | np.int16 |
|
| 67 |
+
| | LC10 | 1 | - | ESA WorldCover 10m v100 product. | np.int16 |
|
| 68 |
+
| LABEL_ | fmask | 1 | - | Fmask4.0 cloud masking. | np.int16 |
|
| 69 |
+
| | QA60 | 1 | - | SEN2 Level-1C cloud mask. | np.int8 |
|
| 70 |
+
| | s2cloudless | 1 | - | sen2cloudless results. | np.int8 |
|
| 71 |
+
| | sen2cor | 1 | - | Scene Classification band. Obtained from SEN2 level 2A. | np.int8 |
|
| 72 |
+
| | cd_fcnn_rgbi | 1 | - | López-Puigdollers et al. results based on RGBI bands. | np.int8 |
|
| 73 |
+
| |cd_fcnn_rgbi_swir| 1 | - | López-Puigdollers et al. results based on RGBISWIR bands. | np.int8 |
|
| 74 |
+
| | kappamask_L1C | 1 | - | KappaMask results using SEN2 level L1C as input. | np.int8 |
|
| 75 |
+
| | kappamask_L2A | 1 | - | KappaMask results using SEN2 level L2A as input. | np.int8 |
|
| 76 |
+
| | manual_hq | 1 | | High-quality pixel-wise manual annotation. | np.int8 |
|
| 77 |
+
| | manual_sc | 1 | | Scribble manual annotation. | np.int8 |
|
| 78 |
+
|
| 79 |
+
<br>
|
| 80 |
+
|
| 81 |
+
|
| 82 |
+
### **np.memmap shape information**
|
| 83 |
+
|
| 84 |
+
<br>
|
| 85 |
+
|
| 86 |
+
**train shape: (3000, 512, 512)**
|
| 87 |
+
<br>
|
| 88 |
+
**val shape: (3000, 512, 512)**
|
| 89 |
+
<br>
|
| 90 |
+
**test shape: (3000, 512, 512)**
|
| 91 |
+
|
| 92 |
+
<br>
|
| 93 |
+
|
| 94 |
+
### **Example**
|
| 95 |
+
|
| 96 |
+
<br>
|
| 97 |
+
|
| 98 |
+
```py
|
| 99 |
+
import numpy as np
|
| 100 |
+
|
| 101 |
+
# Read high-quality train
|
| 102 |
+
train_shape = (3000, 512, 512)
|
| 103 |
+
B4X = np.memmap('train/L1C_B04.dat', dtype='int16', mode='r', shape=train_shape)
|
| 104 |
+
y = np.memmap('train1/manual_hq.dat', dtype='int8', mode='r', shape=train_shape)
|
| 105 |
+
|
| 106 |
+
# Read high-quality val
|
| 107 |
+
val_shape = (3000, 512, 512)
|
| 108 |
+
B4X = np.memmap('val/L1C_B04.dat', dtype='int16', mode='r', shape=val_shape)
|
| 109 |
+
y = np.memmap('train2/manual_hq.dat', dtype='int8', mode='r', shape=val_shape)
|
| 110 |
+
|
| 111 |
+
|
| 112 |
+
# Read high-quality test
|
| 113 |
+
test_shape = (3000, 512, 512)
|
| 114 |
+
B4X = np.memmap('test/L1C_B04.dat', dtype='int16', mode='r', shape=test_shape)
|
| 115 |
+
y = np.memmap('train3/manual_hq.dat', dtype='int8', mode='r', shape=test_shape)
|
| 116 |
+
```
|
| 117 |
+
<br>
|
| 118 |
+
|
| 119 |
+
|
| 120 |
+
This work has been partially supported by the Spanish Ministry of Science and Innovation project
|
| 121 |
+
PID2019-109026RB-I00 (MINECO-ERDF) and the Austrian Space Applications Programme within the
|
| 122 |
+
**[SemantiX project](https://austria-in-space.at/en/projects/2019/semantix.php)**.
|