I have been trying to create an interferogram for deformation analysis with ALOS-2 data in SNAP, but the unwrapped phase products have this ramp in it. I followed the process in S1TBX Stripmap Interferometry with Sentinel-1 Tutorial:
Coregistration
Interferometry (with subtract flat earth phase and topographic phase removal)
Goldstein filter
Phase unwrapping
I used an external DEM since my area of interest is further north than the SRTM coverage. The coregistered image looks good and the coherence looks good. The ALOS-2 images are 14 days apart. I have tried with other ALOS-2 image pairs, and the ramp was present in all but one result. Does anyone know what could be causing this ramp? Is there any processing steps I can use to remove it/improve the results? Could it have anything to do with flat earth phase?
The coherence is shown with all pixels over 0.4 in red.
what is the perpendicular baseline of your image pair?
Another reason for the ramp can be that areas of higher coherence are only fragmented in your image. Not ideal for the unwrapping process, because errors largely add up.
The perpendicular baseline on this image pair is ~70 meters. On the other pairs varies between 138 m and 23 m.
The image pair with baseline of 23 meters looks different from the other results (I made 4 interferograms in total), which are similar to the images I showed in my opening post.
This is the results for the image pair with perpendicular baseline of 23 meters. First is coherence, where coherence >= 0.4 is shown in red. Second image shows unwrapped phase, where blue = 1.8 and red = -7.4.
Could state the elevation too. As @ABraun is emphasizing in several posts here the varied topography present is not easy and trivial to get reliable results. Also you mentioned that the ROI is north above 60 degrees which for me implies presence of snow/ice and not very high coherence I guess.
