Problems with TOPS Split Operator in Graph Builder

Hello fellow STEP-ers and (whipper)SNAP-ers,

I am experiencing a strange problem when trying to automate a processing pipeline for interferometric purposes. I am trying to assemble a graph to automate as much of the process as I can, but an unexpected problem arises with the TOPSAR Split operator. There is no error, but it is not behaving as I expect it to - no sub-swath selection to speak of and apparently the image has 9999 bursts. Check it out:

There is also no demarcation of where my images are supposed to be (indicated by the red circle-ish blob). My two inputs are two SLC images (the respective planned to be master and slave) Side-note: I have already ran the whole process manually but want to now give it a stab at automation. Am I doing something wrong? Has anyone experienced such behavior? Am I asking too much from the Graph Builder’s humble capabilities?

Regards,
Momo

I can’t help you with your question, but I like (whipper)SNAP-ers! :smile:

2 Likes

I usually have this “no subswath” thing, when the graph is freshly opened in Graph builder. I don’t have any serious solution, but it fixes itself after “clicking around”. Sometimes Graph wants to see “Write” box, but usually clicking on “Read” (I have product locations saved in graph) helps. If You (and Graph) check, what to read, Graph somehow clarifies, what to split.

I already encountered this before.

What you need to do is to wait a little bit. Set your other parameters and then come back to your split modules. Normally you should select your IW/EW bands and bursts.

Or you can edit your generated xml file if you saved it. It’s quite simple. For example :

  <node id="TOPSAR-Split">
    <operator>TOPSAR-Split</operator>
    <sources>
      <sourceProduct refid="Read"/>
    </sources>
    <parameters class="com.bc.ceres.binding.dom.XppDomElement">
      <subswath>EW1</subswath>
      <selectedPolarisations>HH</selectedPolarisations>
      <firstBurstIndex>1</firstBurstIndex>
      <lastBurstIndex>8</lastBurstIndex>
      <wktAoi/>
    </parameters>
  </node>
1 Like

Thanks to all for the suggestions. In the end, as unexplainably as it appeared, so did it seem to have resolved itself… must be some kind of momentary bug.

P.S. If I can give any advise though, it would be don’t construct graphs as big as this, I ended up breaking it down to smaller parts.