As many people have observed, the S1 SLC products for a specific area of interest don’t always have perfectly matching bursts. For example, the 1st burst in one product may correspond to the 2nd burst in another product.
The upshot is that we’re left with a smaller common area to coregister and fewer burst overlaps to apply Enhanced Spectral Diversity – i.e., determine and apply a global azimuth shift.
I wonder if/how/when the S1TBX will be able to handle this.
Edit: here’s an example over Vienna (overlap in purple).
I wonder whether there’s some feature in the S1TBX which could be used for extracting, for example, nine contiguous bursts for a given area of interest. And so, if we wanted to create a stack of SLCs, we’d need access to more than one SLC product per acquisition date, so as to ensure a minimum number of bursts is available.
I wonder whether there’s any recommended practice for addressing this issue with the S1 Slice Assembly feature, in combination with the Virtual Stack Processing feature.
So, I assemble them first and then select bursts in the TOPSARSplit op using the wktAoi parameter. However, this approach does not always work. For example, when I coregister these products:
Yes, of course (see attachment). Note that I’m using the ${parameter_name} convention to pass parameters when calling gpt, and that sometimes makes the Graph Builder draw wrong arcs.
The graph builder does not know what to do with the ${parameter_name} parameters. It’s only expecting actual parameter values. It’s different then in the command line where you would pass in these parameters.
I’ve been looking into the xml metadata of those SLCs (zip files as delivered by ESA). It seems that the SliceAssembly op fails due to a missing burst for the products acquired on 20150527. I will contact the scihub support for this, though I can imagine you may want the toolbox to properly handle this exception.
Here’s a visual on how I concluded that there’s a missing burst by loading the geolocationGridPoint xml elements in QGIS:
Date: 20150807 (this one is fine, as the beginning and end of bordering bursts overlap) 1st Slice
