Coherence calculation of TerraSAR-X Staring SpotLight datain SNAP

[radarlove]

I am trying coherence calculation for TerraSAR-X Staring SotLight data.
According to this DLR paper, for spotlight data I need to bring the Doppler spectra to match before I can calcuate the interferogram/corherence.

My graph is like this:

My coherence has a stripe pattern, which runs in parallel to range direction:

If I get rid of the 2 functions and rerun the graph:

The stripes are gone now, but I am lacking cohrence and a large number of small coherent objects I need for furhter analysis:

Here is the comparison:

Her eis my color composite of T1, T2 amplitude on RED and GREEN and the coherence on BLUE.

How can I obtain coherence variant 1, but without the range-parallel stripes? Where is my mistake?

Thanks for looking at it.

[jun_lu]

We have run your graph with the Uluru_SlidingSpotlight TerraSAR-X data we have and the default settings of the operators. Unfortunately we could not reproduce the stripes in the result (see images below). What parameters did you use in your processing?

[radarlove]

Thanks for looking at my issue, jun_lu, but the stripes do not occur in TerraSAR-X High Resolution SpotLight images over Uluruh, Australia, that you’re showing here. My example came from a Staring Spotlight (squinting forward and backward by 2.1 degrees each), an image pair over Rotterdam.

I also tried it with a image InSAR triplet over Schwedt, Germany, which I downloaded here: Sample geospatial Imagery | Radar Constellation Staring SpotLight, Schwedt, Germany

The coherence of pair-1 is fine, in the coherence of pair-2, I see vertical/diagonal stripes.

My model looks like this:

Thanks for looking at the issue.