Datasets:
Tasks:
Other
Modalities:
Image
Size:
100K<n<1M
Tags:
IGN
Environement
Earth Observation
Aerial Lidar
Point Cloud Segmentation
3D Scene Understanding
License:
CharlesGaydon
commited on
Commit
•
93e08ae
1
Parent(s):
f7d92b9
Update README.md
Browse files
README.md
CHANGED
@@ -16,6 +16,10 @@ tags:
|
|
16 |
|
17 |
This repo contains the orthoimageries used to colorize point clouds in the [FRACTAL](https://huggingface.co/datasets/IGNF/FRACTAL) semantic segmentation dataset.
|
18 |
|
19 |
-
They are provided as is, in their original spatial resolution (0.2 m) and band order (near-infrared, red, green, blue).
|
20 |
|
21 |
-
They might be used for quick visual inspections of the FRACTAL's point clouds, or for more advanced use such as multimodal (2D-3D) deep learning training.
|
|
|
|
|
|
|
|
|
|
16 |
|
17 |
This repo contains the orthoimageries used to colorize point clouds in the [FRACTAL](https://huggingface.co/datasets/IGNF/FRACTAL) semantic segmentation dataset.
|
18 |
|
19 |
+
They are provided as is, in their original size (50 x 50 m), spatial resolution (0.2 m, i.e. 250 x 250 pixels) and band order (near-infrared, red, green, blue).
|
20 |
|
21 |
+
They might be used for quick visual inspections of the FRACTAL's point clouds, or for more advanced use such as multimodal (2D-3D) deep learning training.
|
22 |
+
|
23 |
+
Note:
|
24 |
+
- The files are ordered by filename. Inspecting a single zip archive is therefore not representative of the whole dataset.
|
25 |
+
- The band order does not play nice with standard image softwares. Consider reordering band order and disabling the use of transparency in a GIS software like [QGis](https://www.qgis.org/fr/site/).
|