geocode issue

[sunji...@gmail.com]

Hi, Yunjun, 

Thanks for the nice code!

I had a try with Mintpy on 2-frame S-1 data, produced with Gamma software with SRTM 1-arc-second DEM. 

I did small baseline analysis on 8x2 look interferograms, but it seems slow in the processing. In order to accelerate the analysis, I applied a multi-look factor of10-by-10  in the cfg file further. It is smooth and fast. However, the geocoding results are wrong. Please check the attached figure for one of the results. 

I guess the 10x10 average results were not recovered to the original resolution before geocoding, and this leads to the problematic geocoding results. 

Best wishes, 

    Jianbao 

 

 

[Zhang Yunjun]

Hi Jianbao,

The geocoding issue when enabling mintpy.load.x/ystep is a bug, it should be fixed here: https://github.com/insarlab/MintPy/pull/507. Thank you for identifying it!

Yunjun

[sunji...@gmail.com]

Hi, Yunjun, 

Thanks for your updating! 

I noted a possible reason for the segmentation fault during velocity inversion. 

It does not result from limited physical memory available, but due to the wrong version of scipy we used. After updating to the new version, I can do processing with 0.1 billion points!

That's nice!

Best,

     Jianbao