Interferograms processed by stripmapApp are smaller than the same processed by insarApp

[fdelgadodelapuente]

Dear ISCE developers,

I've noticed that ENVISAT, ALOS and COSMO-SkyMed interferograms processed from raw data by stripmapApp.py have a smaller swath (40x40 km2) than the same datasets processed by the deprecated insarApp.py (50x50 km2). Is it due to the different focusing algorithms in the workflows or to some other difference?

Thanks
Francisco

[EJFielding]

Hi Francisco,
Yes, the two processing systems have different focusing algorithms and different parameters. The old insarApp.py used a motion-compensation algorithm to focus the two scenes of raw data as though they were both acquired on a reference track. I think the parameters for the focusing in insarApp extended the swath to attempt to get all the data even if the baseline between the actual and reference track is long. We used to see a lot of ugly bad phase along the edges with the COSMO-SkyMed interferograms processed with insarApp because it was extending the radar pulse compression beyond where there is useful data. I expect there was similar extension in the azimuth direction. You can probably see that the radar amplitude decreases in the near and far range due to the range extension with a partial pulse.
The stripmapApp.py program uses a basic range-Doppler focusing algorithm with the original orbit track and Doppler centroid, similar to the "roi" program in the old ROI_pac package. It probably has the range and azimuth extension parameters set to smaller values because it does not need to process more area to get all the data.

[fdelgadodelapuente]

Thanks Eric for the detailed explanation.

Here is an example of a COSMO-SkyMed Interferogram of Villarrica volcano in southern Chile and you can see that the insarApp swath (dark amplitude and phase) is ~10 km longer in the ground range direction than the stripmapApp swath (on top). There is almost no difference in the azimuth direction.

Do you know how to change the range extension parameters in stripmapApp?

Thanks
Francisco

[EJFielding]

I don't know how to change the range extension parameters in stripmapApp. I think those parameters are not exposed for modification in the stripmapApp.xml. I remember spending some time trying to reduce the range extension in insarApp back around 2014 because the phase at the extreme edges (where the amplitude is very dark) was often some kind of noise that ends up being coherent in interferograms and messes up the phase unwrapping. The way that insarApp and the modules it calls were written, it was not easy to change the extension without a lot of rewriting.

[fdelgadodelapuente]

Thanks Eric. Actually the range extension parameters are not part of the stripmapApp inputs. I'll have to dive into the software internals and try to change them.

Francisco

[EJFielding]

The ISCE2 code is structured in a way that makes changing things like that quite difficult. Not like the old ROI_pac code.

[fdelgadodelapuente]

Do you know a paper where the range extension concept is described in detail?

[EJFielding]

I don't know the SAR signal processing literature well. It is using partial chirp compression where the program uses a portion of the reference chirp to match the signal data past the point where the full chirp is present. A paper on range compression methods should explain this.