Sentinel-1 processing

[Pekka Hurskainen]

Hi,

I have two years time series of S1A and S1B data from my study area, in ascending and descending orbits. I have already done the necessary preprocessing steps to "Level2" in SNAP, so the data can be considered "analysis-ready". Now I would wish to proceed with FORCE, specifically to

1) create a force-compatible level-2 data cube

2) calculate spectral-temporal metrics from these with force-higher-level

I couldn't find much help in the documentation or tutorials how to proceed, only that the input data should be in 16bit signed integer. Any hints?

Pekka

[david frantz]

Hi Pekka,

there is a bit more information in the version history - this is not much, I know..

https://force-eo.readthedocs.io/en/latest/history/v3.html

The Higher Level Processing System is able to process Sentinel-1 SAR data! You can perform all available time series analyses, Spectral Temporal Metrics, compositing etc. as if it would be a spectral index from optical data. Please note however that there is no FORCE module implemented to preprocess the SAR data (any volunteers to integrate this?). The S1 data need to be prepared in a FORCE-compatible format: they need to be in the correct tiling scheme (see e.g. force-cube). The images need to be signed 16bit integers with scaled backscatter in the order of -1000s, nodata value needs to be -9999. The data need to have two bands:

1 VV
2 VH

Four new “sensors” (like LND08 or SEN2A) have been introduced, i.e.

SENSOR Description
S1AIA Sentinel-1A IW Ascending
S1AID Sentinel-1A IW Descending
S1BIA Sentinel-1B IW Ascending
S1BID Sentinel-1B IW Descending

This allows to merge (or keep them separated) data from ascending and descending orbits, and from S1A and S1B. 

Data needs to be named like this: 20180108_LEVEL2_S1AIA_SIG.tif

Does this help?

David

[Pekka Hurskainen]

Hi,

Thanks for these tips David! I was indeed able to process my Sentinel-1 images to Force Level3 TSA, and calculate the spectral-temporal metrics. Here is my workflow in a nutshell:

1) I downloaded all Sentinel-1A + 1B IW Level-1 GRD products from Copernicus Open Access Hub, from my study area in both ascending and descending orbits, for the time period of two years. 

2) I run my preprocessing script in SNAP to the L1 data, to create the level2 data. Output data is in WGS84 Lat/Lon Float32 Geotiff with ~10 m ground resolution (but still in decimal degrees).

3) I made a QGIS model to scale the L2 Float32 data to Int16, using scaling factor 10 000 and -9999 as nodata value (this could also be done with GDAL). The filesize dropped 50%. At this point, the output files were also named according to the naming convention you pointed out, e.g. 20170522_LEVEL2_S1BID_SIG.tif

4) Run force-cube to dice and resample the L2 data into my grid (the same I used for Sentinel-2 data). I used nearest neighbor in resampling to WGS84 UTM, as it preserves the original pixel values more rigorously. 

5) Edit the parameter file for force-higher-level TSA. This was straight-forward as I could modify my existing parameter file for Sentinel-2, with small modifications. 

6) Run force-higher-level to create the spectral-temporal metrics.

The results look very nice visually. The pixels also align nicely with my Sentinel-2 data in the same cube.

So far, I have done this only for my S1B descending data, so I have to still do the same for my S1A data, which are all in ascending orbits. I will have to see if it makes more sense to calculate the spectral-temporal metrics for the whole dataset, or separately for ascending and descending.

As a conclusion, it is very much possible to create a data cube of Sentinel-1 data in Force, and calculate spectral-temporal metrics out of it, but at this point my workflow requires quite many manual steps. I am sure it could also be simplified and improved, for instance the scaling step (3) could be included in the SNAP processing script. 

Best regards

Pekka

[david frantz]

Excellent Pekka!

And thanks very much for sharing your workflow, much appreciated.

Cheers,

David