No Coherence Mask In Final Product

Panji Brotoisworo

unread,

Dec 11, 2020, 10:28:55 AM12/11/20

to MintPy

Hello,

I think the script can run successfully however I have no coherence masking. My spatial coherence seems to be okay, however all the other types of coherence do not look good. Here is my processing environment:

MintPy release version v1.2.3, release date 2020-07-14
Python 3.8 Anaconda
Windows 8.1
SNAP 7 to process Sentinel-1 SAR
IW2 Subswath bursts 6-8

I follow the instructions that were given in the documentation and and follow the example directory given.

I tested mintpy with the sample data that was also using SNAP and it can work, but it does not work with my own dataset for some reason.

The only reason I can think of is maybe I'm already applying a water mask in SNAP, this is so I can speed up the phase unwrapping process with SNAPHU. Perhaps this is the issue? Another issue is that I cannot run plot_smallbaselineApp.sh

The outputs are seen below.

avgSpatialCoh.png

timeseries_ERA5_demErr.png

maskConnComp.png

ifgstack.png

temporalCoherence.png

taal_template.txt

Zhang Yunjun

unread,

Dec 11, 2020, 8:05:30 PM12/11/20

to MintPy

Hi Panji,

It's good that you got the result without runtime error. Your masking looks fine to me, except that the crater lake seems not been masked out, but this should not be a problem.

For a more meaningful small baseline time-series result, here are my suggestions.

1. the spatial coherence is pretty low in my opinion, you might want to apply more looks and stronger filter to increase the coherence

2. temporal coherence makes sense only in a redundant network (check detailed explanation in section 6.3 in my paper if you are interested), so make more interferograms. I usually have at least a redundancy of 3 (3X interferogram for X SAR images) for my own stack processing.

Maybe other SNAP users have more suggestions.

Yunjun

Panji Brotoisworo

unread,

Dec 11, 2020, 9:44:25 PM12/11/20

to MintPy

Thanks for the response Yunjun.

I will try and redo the analysis with a stronger filter and more looks. I think this is using a redundant network with 3X interferograms. Please see attached to see the outputs of SNAP. Maybe I do not have enough? I'm testing this workflow with a few images only before i proceed to analyze the full dataset of 30 radar images.

For reference, I'm also sharing the multilook and filtering parameters I used in SNAP.

ifg_outputs.png

processing_params.png

ifg_dir_structure.png

ifg_network.png

Panji Brotoisworo

unread,

Dec 12, 2020, 7:32:11 AM12/12/20

to MintPy

Sorry, my mistake. I had a step where it was converting the unwrapped phase to LOS displacement. It is working fine now and has a coherence mask. I'm still experimenting with the multilook and filtering parameters to fix my low coherence.

mintpy_output.PNG

Zhang Yunjun

unread,

Dec 13, 2020, 2:41:17 PM12/13/20

to MintPy

I agree that you are using 3X interferograms, except for the last 2 radar images.

Yunjun

Reply all

Reply to author

Forward