UAVSAR reference xml

[ZachHoppinen]

Hello,

I am interested in processing pairs of UAVSAR images (either from an UAVSAR SLC stack or from PolSAR images) to form interferograms. I was hoping to find an appropriate XML file example in the examples/input_files directory but didn't see one in there. The only file was isceappUAVSAR_Stack.xml which seems to go with a polarimetric analysis of ground projected files from the tutorial on the ISCE2-docs (chapter 12). Is there another reference XML file for uavsar files somewhere?

Or possibly the InSAR app doesn't support UAVSAR? I just assumed it did because of ReadMe: "At this time the sensors that are supported are the following: ALOS, ALOS2, COSMO_SKYMED, ENVISAT, ERS, KOMPSAT5, RADARSAT1, RADARSAT2, RISAT1, Sentinel1, TERRASARX, UAVSAR and SAOCOM1A."

Thanks for your help!

[ZachHoppinen]

It seems like based on: https://github.com/isce-framework/isce2/blob/def109815df5f724c5dc653878314577e3505c48/components/isceobj/Sensor/UAVSAR_Stack.py that my reference and activity index may need an: "annotation file" component and a "segment index" component. The annotation file part makes sense but does anyone have any clues as to what segment index is?

[EJFielding]

I assume you are you are trying to process UAVSAR coregistered SLC stack data with ISCE2. You should use the isce2/contrib/stack/stripmapStack workflow that is designed to work with coregistered SLC images.

The Jupyter notebook by @forrestfwilliams https://github.com/forrestfwilliams/UAVSAR_InSAR describes the workflow applied to UAVSAR SLC stack data.

[ZachHoppinen]

Have there been comparisons between the two showing that SNAPHU is doing a better job? Ie should I go back through and increase my multilooking to use SNAPHU? Is there an advantage to that over just using icu?

I am using 8 looks in azimuth 2 in range, just based on Forrest's notebook so I will increase to 3x12 and see if I can get the snaphu to run if neccessary/useful.

[EJFielding]

Umm, there are many ways ICU and SNAPHU differ. I don't have time to explain. If you got good results with ICU, you can roll with that.

[ZachHoppinen]

Sounds good. I will multi-look and compare the two and see if there is a difference. Thanks again for your time Eric!

[ZachHoppinen]

So after running through this process the unwrapped interferograms I generated have almost exclusively values above zero. Is there something about the interpretation of these ICU unwrapped phase values I am missing that would make the unwrapped phase pixels all above zero? Or is there some step that I missed that caused this?

[EJFielding]

InSAR phases are always relative. Different unwrapping programs choose different places to put the "zero" for the unwrapped phase. I think SNAPHU uses the upper left corner and ICU uses some kind of average of the top of the image. Either way, there is always an unknown bias constant for the InSAR phase. You need other information to determine what is a more accurate zero level and what constant to add or subtract.