search

umr-lops / L2A-wind-direction-processor

Python package to generate Level-2A wind direction files from Sentinel-1 TOPS product based on low frequency texture analysis.
MIT License
4 stars 0 forks source link

clarify input Level-1B tiles size #18

Closed agrouaze closed 5 months ago

agrouaze commented 5 months ago

default values for sigma0 patches in original network developed by @rmarquarlops are: 44x44 pixels. These 44x44 pixels matrices are achieved using interface.py on a 17.6km² tiles at 400 m resolution in azimuth and range:

def get_low_res_tiles_from_L1BSLC(file_path, xspectra = 'intra', posting = {'sample':400,'line':400},
                                  tile_width = {'sample':17600.,'line':17600.}, window='GAUSSIAN', **kwargs):

but it seems that using Level-1B generated with a 17600m² size is leading to have sigma0 patches with NaN on the edges:

image While we can get 44x44 sigma0 matrix starting from Level-1B produced on 17.7km² tiles without any NaN: image

Question for @rmarquarlops : Should we start from Level-1B product with larger tiles (i.e. 17.7km² instead of 17.6km²)? Having NaN in the sigma0 matrix does not prevent to do the prediction but the results are completely different compare to matrix without NaN.

agrouaze commented 5 months ago

Found the answer to this question: Frederic Nouguier wrote:

Tu verras qu'il y a deux trois options dans la fonction qui te permettront d'éventuellement changer la taille et le posting des tiles de sortie. Si les L1B ont été générés sur des tiles de 2km x 2km, tu ne pourras pas demander des tiles de 17.6 km. Tu ne peut extraire que des tiles qui sont plus petites que celles présentes dans le L1B. Pour ton appli, il faudra donc que tu prennes les L1B traités avec des vignettes de 20km x 20 km

So it means that the Level-1B tiling is just a way to crop the SLC. image