datagrid and sbas.find_pairs couldn't imported about pygmtsar

[AlexeyPechnikov]

The colormap available in modern Python versions like to 3.10. In the basic PyGMTSAR examples ‘turbo’ colormap is not used because it’s not available on Google Colab.

[AlexeyPechnikov]

What PyGMTSAR version do you use? Check it as below:

pip3 show pygmtsar
Name: pygmtsar
Version: 2022.12.1

[Meysam-Amiri]

I noticed that my PyGMTSAR version was old and I have updated it with python3 setup.py install using the directory of pygmtsar. Thanks for your help.

[Meysam-Amiri]

In plotting it seems that I don't have the turbo colormap. I couldn't figure out how to add it! I have changed it to rainbow.
error_cmap.txt

[Meysam-Amiri]

@mobigroup, I have processed my data thanks to Yamchi_dam notebooks. How can I improve the results so they wouldn't be like separate pixels in the last sbas map? Does it depend on the resolution that we define or chunk_size?
One more question, How can I get the time series of a point? Is there any example in notebooks?
Here are my notebooks:
https://drive.google.com/drive/folders/1hpCkLDqMV_lqgYUAnzfJAMn-OiiBUNUl?usp=sharing
Thanks a lot.

[AlexeyPechnikov]

@Meysam-Amiri I think you'd use mask with zeroed area for y<5000 instead of composite_mask for unwrapping. On your results we see some phase discontinuity below y=5000 due to decorrelation:

Pixels are sparse because of correlation stack mask applied (corr_mask) and we are able 1) exclude it for SBAS processing or 2) fill the gaps after unwrapping using nearest interpolation. For your area the both ways look suitable.

We can define the manual mask like to

manual_mask = (corrs[0].where(corrs.y >= 5000)>=0).astype(bool).persist()

and use it for unwrapping as

sbas.unwrap_parallel(mask=manual_mask, conf=conf, n_jobs=...)