[zahrasadeghi]

Hi,
I’ve done a QPSI processing using my 2.5 year S1 images, and only found a few points around my small case study, at the moment I am working on different things (working around thresholds etc) to help to increase the density. Could you please let me know if you have any specific advice or suggestions on bellow:

• MT adaptive mask: I selected AD with 5 *5 win size, 0.05 for signif level and 4 pixel for connectivity, and called this in the InSAR params module, how changing these values can help to increase density?
• Insar params module: selecting pre-filtering or post-filtering for output coherence can affect the density? I selected sparse LS for unwrapping as it sounds working better than standard,
• Interferograms processing: is this step mandatory for QPSI for a full graph? It’s very time-consuming for a long time series of data, is it enough to only run full graph coherence estimation while selecting MT adaptive filter in Insar params and skip interferogram procesing?
• Do you recommend Amp.Stab.Index 1-S/M FOR APS estimation/ sparse point processing and temporal coherence for TS module for point selection?

Many thanks,
Zahra

[periz]

Hi Zahra,
firstly, few general comments:
* there is no perfect solution that works in all situations. you have to choose the proper solution given your particular situation
* the secret of the best InSAR expert is being able to look at and read the data. only afterwards one can properly choose the processing options
* we try to teach looking at the data during the Sarproz course.
Given the above general directions,
I can add that if you do not process interferograms in advance, when the sw tries to read the phase from interferograms and it does not find them, it will give a message that says that the interferometric phase will be estimated using a fixed boxcar filter.
The full graph QPS approach is energy and time consuming. It is your call to decide whether it is worth using it or not. But do not treat it as the magic box that gives you solutions. As stated before, use it only after you have already understood what your data look like and what your data can tell you.

[Author]

Posts