Artefacts after Terrain Flattering of Sentinel-1 with COP DEM

[Postalstar]

Hi all. I’m running Terrain Flattering and Terrain Correction of Sentinel-1 GRD products in SNAP 11.0.1 and 12 and experiencing artefacts similar to the ones, described in Artefacts in CopDEM 30m Auto Download - Microwave Toolbox / Problem Reports - STEP Forum, but there are noticeable differences and specifics.

First of all, only vertical artefacts appear (as in previous topic at the DEM tiles boundaries). No horizontal distortions are noticed.

Secondly, they appear only higher that 70 degrees north latitude. As far as I seen, everything to the south of 70 degrees is fine. Maybe artefacts reappear after 70 degrees south latitude, didn’t checked it.

Third, it happens with both 90m Copernicus DEM product and 30m one. With 30 m its a bit less noticeable, but still there. Changing resampling method also helps a little bit, but does not resolve the problem.

Forth, I checked generated DEM in Terrain Correction procedure and it present there.

It especially noticeable in more mountainous area, but on flat terrain you still can see it.

It also have to be noted that as in mentioned above topic, It was not an issue in SNAP version 8.

So I guess the solution, that was implemented in SNAP 10 patches did not affected mosaicking and resampling for DEM tiles of these Arctic areas, maybe because northern than 70 degrees pixel size changes and current version does not takes it onto account.

[jun_lu]

Could provide the product name, processing steps and parameters used so that we can reproduce the problem? Thank you.

[Postalstar]

Hi, yes.
It happens with any product higher than 70 degrees north, for example S1B_IW_GRDH_1SDV_20211008T220352_20211008T220417_029050_03776A_CD9B.
Processing undergoes steps within these 2 graphs.
orb_tnr_cal_beta_gpt_VV_HH.xml (2.9 KB)
flattening_terr_corr_db_gpt_NEWSNAP_30m.xml (5.3 KB)

[jun_lu]

A Jira ticket (Jira) has been created to track the problem. We will look into the issue. Thank you.

[jun_lu]

The problem has been fixed. The fix should be available in the next release. Thank you.

[Postalstar]

Thank you very much for quick response. If it possible, can you please tell approximate time window for this next release? like couple of months?

[jun_lu]

Probably in a couple of months.

[Postalstar]

Hi, I checked new release of SNAP and tried to process several scenes across the Arctic. As you can see on screenshots, situation with artefacts got better, but they are still noticeable, unlike in much older version of SNAP 8 or 9 (where they are not present at all). Is it possible to improve current state of processing? Scene ID of provided example is S1C_IW_GRDH_1SDH_20250504T092254_20250504T092323_002179_004A14_D221

[diana_harosa]

@lveci could you take a look at this issue?

Thank you!

[Postalstar]

Hi, are there any updates? Is it possible to fix this issue and if yes, what is approximate time window for this?

[Postalstar]

Hi, I tried new release of Microwave toolbox 13.02. Everything got even worse. These DEM tiles stitching artefact is still present, but just terrain corrected now.

Moreover - now scene is not processed fully, it just stops processing at certain point with disappearing triangles. I tried it with several scenes and results are all the same. Scene ID of provided example is S1C_IW_GRDH_1SDH_20250504T092254_20250504T092323_002179_004A14_D221

And if first type of artefacts (on DEM tiles borders) appears as previously northern from 70 degrees north (and probably same for south), second (incomplete processing) appears elsewhere.

[s.savastano]

Hi @Postalstar ,

thanks to have posted this. I’m having similar issue in the Terrain Flattening step. After the last update (13.0.2) the TF is applied not to the whole scene but on portion of it.

in my case my input image is
S1A_IW_GRDH_1SDV_20200917T175447_20200917T175512_034402_040042_FF4D.
I tried to change from Cop dem to SRTM 1Sec but the problem is still there.
Any help for this @jun_lu @diana_harosa ?

Thanks in advance
S Savastano

PS Just in addition, even if I don’t do the TF step, a similar issue happens during the DEM coregistration creating in output a black image as slaves

[diana_harosa]

Jira ticket SNAP-4148 created.