Linear artifact in Sentinel-1 Orthorectified images in EO Browser

There’s a strange linear artifact in the orthorectifed S1 images in the EO Browser, which looks like a pixellated edge of something.

Is it caused by the orthorectification? If so, what feature is it, an outdated water surface edge?

Page 1 / 1

Hm, I am not sure if I see what you see as I do not notice anything specifically strange.

In case the issue is only with orthorectification, the likely culprit is digital elevation model, which is far from perfect. You should be able to notice similar patterns in it, e.g. here:

apps.sentinel-hub.com
89e18423-6a2f-40bb-bf43-6bd846d6fa6c?showLogo=false&service=WMS&request=GetMap&layers=DEM_COLOR&styles=&format=image%2Fjpeg&transparent=false&version=1.1.1&name=sentinel2&height=400&width=400&maxcc=100&priority=mostRecent&gain=1&gamma=1&ATMFILTER=none&time=2013-01-01%2F2020-10-18&srs=EPSG%3A3857&evalsource=DEM&PREVIEW=3&bbox=3935852,3711275,3940744,3716167

89e18423-6a2f-40bb-bf43-6bd846d6fa6c?showLogo=false&service=WMS&request=GetMap&layers=DEM_COLOR&styles=&format=image%2Fjpeg&transparent=false&version=1.1.1&name=sentinel2&height=400&width=400&maxcc=100&priority=mostRecent&gain=1&gamma=1&ATMFILTER=none&time=2013-01-01%2F2020-10-18&srs=EPSG%3A3857&evalsource=DEM&PREVIEW=3&bbox=3935852,3711275,3940744,3716167

Sentinel-hub Playground

DEM imagery taken on October 18, 2020

Yes, as I suspected it is the DEM which is at fault.

Which DEM product is it? it has many problems

Mapzen’s one:
https://registry.opendata.aws/terrain-tiles/

it is not best indeed, although in parts much better than SRTM30. Once we find some other one available on the cloud infrastructure, we will consider switching. Or perhaps make it an option.