Datasets:

ArXiv:
DOI:
License:
neurograce commited on
Commit
86f5129
·
verified ·
1 Parent(s): d84aba9

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +21 -18
README.md CHANGED
@@ -1,18 +1,21 @@
1
- # Remote Sensing Dataset: Substation Dataset
2
-
3
- ## Description
4
-
5
- The first dataset is curated by TransitionZero and sourced from publicly available data repositories, including OpenSreetMap (OSMF) and Copernicus Sentinel data. The dataset consists of Sentinel-2 images from 27k+ locations; the task is to segment power-substations, which appear in the majority of locations in the dataset. Most locations have 4-5 images taken at different timepoints (i.e., revisits) and each image is of dimension 228x228 pixels. Each image has 13 spectral bands and each band has been linearly interpolated to a spatial resolution of 10m. Lastly, there is one ground truth mask for each location.
6
-
7
- ### Key Features
8
-
9
- - **Source:** OpenSreetMap (OSMF) and Copernicus Sentinel data
10
- - **Resolution:** 10m per pixel
11
- - **Bands:** 13 Sentinel-2 Bands
12
- - **Size:** Approximately 70GB
13
-
14
- We utilize this dataset in this [project](https://arxiv.org/abs/2409.17363). In this work, we focus on an applied research question of relevance to climate change mitigation -- power substation segmentation -- that is representative of applied uses of pre-trained models more generally. Through extensive tests of different multi-temporal input schemes across diverse model architectures, we find that fusing representations from multiple revisits in the model latent space is superior to other methods of using revisits, including as a form of data augmentation. We also find that a SWIN Transformer-based architecture performs better than U-nets and ViT-based models.
15
-
16
- ---
17
- license: apache-2.0
18
- ---
 
 
 
 
1
+ ---
2
+ license: cc-by-4.0
3
+ ---
4
+ # Remote Sensing Dataset: Substation Dataset
5
+
6
+ ## Description
7
+
8
+ The first dataset is curated by TransitionZero and sourced from publicly available data repositories, including OpenSreetMap (OSMF) and Copernicus Sentinel data. The dataset consists of Sentinel-2 images from 27k+ locations; the task is to segment power-substations, which appear in the majority of locations in the dataset. Most locations have 4-5 images taken at different timepoints (i.e., revisits) and each image is of dimension 228x228 pixels. Each image has 13 spectral bands and each band has been linearly interpolated to a spatial resolution of 10m. Lastly, there is one ground truth mask for each location.
9
+
10
+ ### Key Features
11
+
12
+ - **Source:** OpenSreetMap (OSMF) and Copernicus Sentinel data
13
+ - **Resolution:** 10m per pixel
14
+ - **Bands:** 13 Sentinel-2 Bands
15
+ - **Size:** Approximately 70GB
16
+
17
+ We utilize this dataset in this [project](https://arxiv.org/abs/2409.17363). In this work, we focus on an applied research question of relevance to climate change mitigation -- power substation segmentation -- that is representative of applied uses of pre-trained models more generally. Through extensive tests of different multi-temporal input schemes across diverse model architectures, we find that fusing representations from multiple revisits in the model latent space is superior to other methods of using revisits, including as a form of data augmentation. We also find that a SWIN Transformer-based architecture performs better than U-nets and ViT-based models.
18
+
19
+ ---
20
+ license: apache-2.0
21
+ ---