NGonthier commited on
Commit
5a22d70
1 Parent(s): e328944

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +62 -0
README.md CHANGED
@@ -1,3 +1,65 @@
1
  ---
2
  license: etalab-2.0
 
 
 
 
 
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: etalab-2.0
3
+ task_categories:
4
+ - image-classification
5
+ - image-segmentation
6
+ tags:
7
+ - climate
8
+ - remote sensing
9
+ - Agricultural
10
+ size_categories:
11
+ - 1K<n<10K
12
  ---
13
+
14
+
15
+ # 🌱 PASTIS-HD 🌿 Panoptic Agricultural Satellite TIme Series : optical time series, radar time series and very high resolution image
16
+
17
+ [PASTIS](https://github.com/VSainteuf/pastis-benchmark) is a benchmark dataset for panoptic and semantic segmentation of agricultural parcels from satellite time series.
18
+ It contains 2,433 patches within the French metropolitan territory with panoptic annotations (instance index + semantic label for each pixel).
19
+ Each patch is a Sentinel-2 multispectral image time series of variable lentgh.
20
+
21
+ This dataset have been extended in 2021 with aligned radar Sentinel-1 observations for all 2433 patches.
22
+ For each patch, it constains approximately 70 observations of Sentinel-1 in ascending orbit, and 70 observations in descending orbit.
23
+
24
+ We extend PASTIS with aligned very high resolution satellite images from SPOT 6-7 constellation for all 2433 patches in addition to the Sentinel-1 and 2 time series.
25
+ The image are resampled to a 1m resolution and converted to 8 bits.
26
+ This enhancement significantly improves the dataset's spatial content, providing more granular information for agricultural parcel segmentation.
27
+ PASTIS-HD can be used to evaluate multi-modal fusion methods (with optical time series, radar time series and VHR images) for parcel-based classification, semantic segmentation, and panoptic segmentation.
28
+
29
+ The SPOT images are opendata thanks to the Dataterra Dinamis initiative in the case of the ["Couverture France DINAMIS"](https://dinamis.data-terra.org/opendata/) program.
30
+
31
+
32
+ - **Dataset in numbers**
33
+
34
+ 🛰️ Sentinel 2 | 🛰️ Sentinel 1 | 🛰️ **SPOT 6-7 VHR** | 🗻 Annotations
35
+ :-------------------------------------------- | :-------------------------------------------------- | :------------------------------| :------------------------------
36
+ ➡️ 2,433 time series | ➡️ 2 time 2,433 time series | ➡️ **2,433 images** | 124,422 individual parcels
37
+ ➡️ 10m / pixel | ➡️ 10m / pixel | ➡️ **1m / pixel** | covers ~4,000 km²
38
+ ➡️ 128x128 pixels / images | ➡️ 128x128 pixels / images | ➡️ **1280x1280 pixels / images** | over 2B pixels
39
+ ➡️ 38-61 acquisitions / series | ➡️ ~ 70 acquisitions / series | ➡️ **One observation** | 18 crop types
40
+ ➡️ 10 spectral bands |➡️ 2 spectral bands | ➡️ **3 spectral bands** |
41
+
42
+
43
+ ## References
44
+ If you use PASTIS please cite the [related paper](https://arxiv.org/abs/2107.07933):
45
+ ```
46
+ @article{garnot2021panoptic,
47
+ title={Panoptic Segmentation of Satellite Image Time Series
48
+ with Convolutional Temporal Attention Networks},
49
+ author={Sainte Fare Garnot, Vivien and Landrieu, Loic },
50
+ journal={ICCV},
51
+ year={2021}
52
+ }
53
+ ```
54
+
55
+ For the PASTIS-R optical-radar fusion dataset, please also cite [this paper](https://arxiv.org/abs/2112.07558v1):
56
+ ```
57
+ @article{garnot2021mmfusion,
58
+ title = {Multi-modal temporal attention models for crop mapping from satellite time series},
59
+ journal = {ISPRS Journal of Photogrammetry and Remote Sensing},
60
+ year = {2022},
61
+ doi = {https://doi.org/10.1016/j.isprsjprs.2022.03.012},
62
+ author = {Vivien {Sainte Fare Garnot} and Loic Landrieu and Nesrine Chehata},
63
+ }
64
+ ```
65
+