Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -16,12 +16,6 @@ size_categories:
|
|
| 16 |
CrisisLandMark is a large-scale corpus composed of Sentinel-1 and Sentinel-2 data enriched with textual and geospatial annotations
|
| 17 |
designed for LULC and crisis management applications. The primary purpose is to help design text-to-image retrieval pipelines.
|
| 18 |
|
| 19 |
-
## Dataset Details
|
| 20 |
-
|
| 21 |
-
<!-- Provide a longer summary of what this dataset is. -->
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
- **Curated by:** Daniele Rege Cambrin
|
| 26 |
- **License:** Creative Commons Attribution Non Commercial 4.0
|
| 27 |
- **Repository:** [GitHub](https://github.com/DarthReca/closp)
|
|
@@ -31,17 +25,47 @@ designed for LULC and crisis management applications. The primary purpose is to
|
|
| 31 |
|
| 32 |
<!-- This section provides a description of the dataset fields, and additional information about the dataset structure such as criteria used to create the splits, relationships between data points, etc. -->
|
| 33 |
|
| 34 |
-
[
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 35 |
|
| 36 |
## Source Data
|
| 37 |
|
| 38 |
<!-- This section describes the source data (e.g. news text and headlines, social media posts, translated sentences, ...). -->
|
|
|
|
| 39 |
|
| 40 |
-
|
|
|
|
|
|
|
|
|
|
| 41 |
|
| 42 |
### Annotations
|
| 43 |
|
| 44 |
<!-- If the dataset contains annotations which are not part of the initial data collection, use this section to describe them. -->
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 45 |
|
| 46 |
## Citation
|
| 47 |
|
|
|
|
| 16 |
CrisisLandMark is a large-scale corpus composed of Sentinel-1 and Sentinel-2 data enriched with textual and geospatial annotations
|
| 17 |
designed for LULC and crisis management applications. The primary purpose is to help design text-to-image retrieval pipelines.
|
| 18 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
- **Curated by:** Daniele Rege Cambrin
|
| 20 |
- **License:** Creative Commons Attribution Non Commercial 4.0
|
| 21 |
- **Repository:** [GitHub](https://github.com/DarthReca/closp)
|
|
|
|
| 25 |
|
| 26 |
<!-- This section provides a description of the dataset fields, and additional information about the dataset structure such as criteria used to create the splits, relationships between data points, etc. -->
|
| 27 |
|
| 28 |
+
The *crisislandmark.h5* file can be used to access all the keys. Since it contains only [ExternalLinks](https://docs.h5py.org/en/stable/high/group.html#external-links), the other files are necessary.
|
| 29 |
+
Each key contains the following matrices:
|
| 30 |
+
|
| 31 |
+
| Key | Shape | Data Type | Description |
|
| 32 |
+
| :------- | :-------------- | :-------- | :---------------------------------------------------- |
|
| 33 |
+
| `image` | `(B, 120, 120)` | `float32` | Contains the B bands of the image from S1 or S2. |
|
| 34 |
+
| `coords` | `(2, 120, 120)` | `float32` | Contains the coordinates for each pixel of the image. |
|
| 35 |
+
|
| 36 |
+
and the following attributes:
|
| 37 |
+
|
| 38 |
+
| Key | Shape | Data Type | Description |
|
| 39 |
+
| :---------- | :---- | :---------- | :----------------------------------------------------- |
|
| 40 |
+
| `crs` | `(1)` | `float32` | Contains the CRS for the coords matrix. |
|
| 41 |
+
| `timestamp` | `(1)` | `float32` | Contains the associated timestamp if available. |
|
| 42 |
+
| `labels` | `(L)` | `list[str]` | Contains the list of labels associated with the image. |
|
| 43 |
+
|
| 44 |
+
The *metadata.parquet* files contain, for each key, the associated split and the labels from the original source (either CLC or DW).
|
| 45 |
+
|
| 46 |
+
**NOTE**: remember to install *hdf5pluging* package to work with the dataset.
|
| 47 |
|
| 48 |
## Source Data
|
| 49 |
|
| 50 |
<!-- This section describes the source data (e.g. news text and headlines, social media posts, translated sentences, ...). -->
|
| 51 |
+
The CrisisLandMark dataset is composed of over 647,000 satellite images from two primary sources: Sentinel-1 and Sentinel-2 missions.
|
| 52 |
|
| 53 |
+
- Sentinel-1: These are Synthetic Aperture Radar (SAR) Ground Range Detected (GRD) products, which provide all-weather, day-and-night observations. The data includes two polarization channels: Vertical transmit-Vertical receive (VV) and Vertical transmit-Horizontal receive (VH).
|
| 54 |
+
- Sentinel-2: These are Level-2A (L2A) surface reflectance optical products, providing data across 12 spectral bands from ultra-blue to short-wave infrared.
|
| 55 |
+
|
| 56 |
+
The images are sourced from five existing datasets: re-BEN, CaBuAr, QuakeSet, MMFlood, and Sen12Flood. All images have been processed to a uniform spatial resolution of 10 meters and divided into 120x120 pixel patches.
|
| 57 |
|
| 58 |
### Annotations
|
| 59 |
|
| 60 |
<!-- If the dataset contains annotations which are not part of the initial data collection, use this section to describe them. -->
|
| 61 |
+
The dataset features a comprehensive set of structured textual and geospatial annotations created specifically for this work.
|
| 62 |
+
The annotations were generated by harmonizing established land cover classification systems and crisis-specific tags.
|
| 63 |
+
|
| 64 |
+
- For general land cover scenes, annotations were derived from the detailed CORINE Land Cover (CLC) system.
|
| 65 |
+
- For images related to crisis events, LULC annotations were generated by querying the near-real-time global Dynamic World (DW) system.
|
| 66 |
+
- For images sourced from crisis-specific datasets (CaBuAr, MMFlood, Sen12Flood, Quakeset), the original event tags such as "wildfire", "flooding", and "earthquake" were retained and incorporated.
|
| 67 |
+
|
| 68 |
+
*NOTE*: In *mapping.py*, you can find the mapping between CLC and DW used in the original settings.
|
| 69 |
|
| 70 |
## Citation
|
| 71 |
|