Interferogram from Alos palsar 1.1 data

I am extremely thankful to both of you for all your efforts @ABraun, @mdelgado!! I am trying a new stack altogether.

the lines are not really missing, it is rather the pattern itself that looks like this.
I also checked it and the interferogram looks the same with (and without) topographic phase.
I will check later if flat-earth removal makes a difference.

1 Like

Still… missing lines or not, it is important to understand:

  1. if the resampled images are ok
  2. if the complex multiplication of master and slave (the pure interferogram) is ok
  3. if the Flat-Earth removal introduces the weird patterns

Thanks again for doing that! I guess that this understanding is needed to really see how to solve it or if it is any possible bug inside the software

1 Like

Well, I was working with images and this happened:


GC overhead exceeded error message

you ran out of RAM.

https://forum.step.esa.int/search?q=gc%20overhead

The strong trend in phase is more likely due to ionospheric effects, which are about 16 times stronger at L-band than C-band. The orbits for the ALOS satellite are generally good.

The latest version of the ISCE software includes an option for estimating and applying an ionospheric correction.

I could try running this processing in the ISCE software, if @ABraun can post the ALOS granule IDs.

I suspect the strong phase ramp could be due to ionospheric effects, which are stronger at low and high latitudes.

1 Like
  • ALPSRP077840420
  • ALPSRP084550420

Thank you for offering!

Did you find them on the ESA site or the NASA ASF site?

ASF. This is why I expected the orbit files as a potential error source at first.

Dear all,

I have also downloaded the images from ASF and this is what I have got after processing them (no deskewing applied)

From right to left: 1)coherence ; 2) Pure interferogram (Ref phase and Topo included) 3) Interferogram after Ref Phase removal (Flat-Earth removed) and; 4) Interferogram after Ellipsoid and Topo Phase removal.
There is almost no difference between 3 and 4, that could be due to low topography variation as it is a flat area (delta)

Hence, these diagonal fringes either:

  1. are due to wrong orbit informaiton
  2. wrong ellipsoid removal

I have always understood that ALOS-1 had good orbit information and it was not needed to do any orbit modification and hence I would say that SNAP does something weird during the ellipoid removal when using ALOS-1 PALSAR data. Can any of you check this out @lveci or @marpet?

It could be also that in this particular image pair anything else happened…

Just wait for @EJFielding results to understand whether this is a software-related effect . I look forward to it.

This is what I got from DIAPASON applied to a Sentinel-1 pair:

Inteferogram (no unwrapping)

Coherence

You see a few fringes but not thoughout the whole area. Coherence is low

Screenshot please? I cannot see it :frowning:

I updated the post.

1 Like

I think I have finally got it!
You really need to change the standard parameters for the Flat-Earth removal step, selecting maximum polynomial degree and points, and the result is as the one in the figure:

This figure shows: multilooked HH coherence (right) and multilooked phase (left) after Flat-Earth removal

Still clearly visible a trend on the phase due to as @EJFielding mentioned, ionospheric artifact strongly present at lower frequencies as L-band

1 Like

Yes! This now looks like classic ionospheric waves. These short spatial-scale waves are difficult to correct in ALOS data, especially the low bandwidth FBD data that we have for this pair.

I will try the ISCE processing a little later today.

1 Like

Thank you for your response, @ABraun. I understand that I:

Further, I looked into the link you sent in this post:

and have understood the problem.

Dear @ABraun, @mdelgado, and @EJFielding , thank you for taking out time to analyse the problem I have been facing and helping me out in such an in depth manner. I shall forever be indebted to you all. :slightly_smiling_face: I understand that this has been a strenuous process but I am glad that you have found an answer to the problem and made my job easier! :smile:

But here I would like to ask a few questions. At first @ABraun, could you please tell me about DIAPSON processing? Secondly, @mdelgado what was the polynomial degree and the number of GCPS that you chose while working with the images?

I look forward for your response.

Dear @ABraun, I got the answer to my question in the post above. :slight_smile:

I ran the interferogram with the ISCE application stripmapApp, and the coherence looks very good everywhere except in the corner where there seems to be some agriculture. Here is the wrapped phase in radar coordinates.

Here is the wrapped phase geocoded.

The ionospheric correction estimate did not work very well due to the extensive water areas, but I think the remaining ramp of phase across the scene, about one fringe is likely ionosphere.

1 Like