Scaling up Sentinel-1 Interferometry?

Microwave Toolbox
1

I have followed various tutorials on Sentinel-1 interferometry, including this ESA document and this forum post. I can produce DEMs from Sentinel-1 using the SNAP GUI.

I am now looking for advice for how to scale this process. i.e process S1 products for multiple sites / time points. I would be grateful for any pointers. Specifically:

  1. I tried the SNAP graph builder, in the hope that I could use it with Batch Processing. Re. S1 TOPS Split, how can you automatically select bursts over your AOI? Is there a workaround, AFAIK S1 SLC cannot be subset?

  2. I do not appear to have an option to add Snaphu Unwrapping to my graph, I only have Snaphu Export / Import. Is the graph process different?

  3. If mass processing is not possible via SNAP graphs, can it be done using the Snappy API?

  4. Are other APIs more suited to my use case, perhaps PyGMTSAR?

Many thanks,

Jay

2

Certainly, PyGMTSAR (Python InSAR) is a fully programmable environment that can handle processing for both small and large areas. For large area analysis, you have several options. If you have a set of sequential scenes, you can stitch them together and process them as a single raster (using tiled SNAPHU unwrapping). Alternatively, you can opt for burst-based processing, where individual bursts are processed separately and then merged in geographic coordinates (use the burst downloading feature to download only the necessary bursts). A Sentinel-1 DEM generation use case is available at https://InSAR.dev. Additionally, PyGMTSAR provides interactive 2D and 3D maps right out of the box.