LSD4WSD : An Open Dataset for Wet Snow Detection with SAR Data and Physical Labelling

Citation Author(s):
Matthieu
Gallet
LISTIC, University Savoie Mont Blanc
Abdourrahmane
Atto
LISTIC, University Savoie Mont Blanc
Fatima
Karbou
CEN, Centre National de Recherches Météorologiques
Emmanuel
Trouvé
LISTIC, University Savoie Mont Blanc
Submitted by:
matthieu gallet
Last updated:
Thu, 07/06/2023 - 10:35
DOI:
10.21227/b7eb-pj47
Data Format:
Research Article Link:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

LSD4WSD: Learning SAR Dataset for Wet Snow Detection.

The dataset can be found at : https://zenodo.org/record/8111485

The aim of this dataset is to provide a basis for automatic learning to detect wet snow.
It is based on Sentinel-1 SAR satellite images acquired between August 2020 and August 2021 over the French Alps. It consists of 487157 samples of size 16 by 16 by 9 for training and 3668 for testing. For each sample, the associated label is obtained using the Crocus physical model.

The 9 channels are in the following order:

 

  • Sentinel-1 polarimetric channels: VV, VH and the combination C: VV/VH,
  • Topographical features: altitude, orientation, slope
  • Polarimetric ratio with a reference summer image: VV/VVref, VH/VHref, C/Cref
Instructions: 

The utils.py file provides the function for opening the hdf5 file and displays the dataset characteristics.

Below is the associated command line:

>>> python3 utils.py --dirpath ./

The complete processing chain will be added in future versions at the following Github address.