StaMPS-Visualizer, SNAP-StaMPS Workflow

[Jennifer]

thanks a lot @thho

[meli7011]

Good afternoon, could you help me in this step please, I’m new to matlab, after applying ‘ps_plot (’ v-do ‘,’ ts ‘)’ I get this window, how do I select a reference point to export a CSV? .

[thho]

you have to define the ref point before plotting, see the StaMPS manual for that, page 32 if I recall right.

then when you have this window, select the radius in the field under the plot, to include all points make it large. then click in the image where the radius should be applied. The points within the radius are then used to build some objects. in the Visualizer app in the Manual tab is matlab code to extract those objects to a csv file.

TL;DR

• read the StaMPS manual page 32 to set ref point
• read the Manual tab of the Visualizer app to export to csv

BTW, nice study location

[CFEgildan71]

Hello Thho

I’ve been following the steps in manual for make subset of my interest area, I would like to make subset with kml file but I have error installing “rgdal” and “rgeos” packages.

I tried typing "insatall.package(“rgdal”) and clicking on packages tab by gui but I cannot had successful
¿Cloud you help me, please?

[meli7011]

Thank you very much, seriously! Thanks to all your posts I was able to get here being very new on the subject, I hope the questions are not a nuisance.

[thho]

@CFEgildan71, I assume you are using a Unix OS, then open a terminal and do:

sudo apt install libgeos-dev
after that open RStudio again
install.packages('rgeos')

should do.

[hm1u16]

@thho I have followed the code instructions as specified in the manual however I am running into an issue on the line export_res = [lon2 lat2 disp ts]
Error using horzcat
Dimensions of arrays being concatenated are not consistent.

Looking in the workspace the items have the following dimensions:
lon2 = 1542 x 1
lat2 = 1542 x1
disp = 1545 x 1
ts = 9 x 1

How would I go around solving this?

[thho]

Hi @hm1u16, have a look at the posts around this one…sounds familiar, if the error remains, ask again here.

StaMPS-Visualizer, SNAP-StaMPS Workflow

Dear @falahfakhri, this is a known error which occurs when there is a high density or amount of PS. Please have a look into the manual page of the Visualizer, there you can find a documented workaround to avoid the error. You also have to be sure, that you user the right export script for your stamps version, since there have been changes in stamps v4.x… All this is also documented in the manual page within the application. I hope that helps you. Cheer

[Wilhelm_Berrocal]

If I remember correctly, you have to give a radius big enough to get all your points selected, by doing that the dimensions are consistent and you can proceed.

[s6039642]

Hi Everyone.
I tried the code in R which comes with the Shiny Visualizer to subset my data in R. In order to visualize it in Shiny Visualizer, however, it fails. Due to availability of ‘rgdal’ and ‘rgeos’ packages in R version 3.4.4. I updated R to the latest version 3.6.2. It still tells me that the packages are not available.
My question is this:
Has anyone faced similar problems with R version 3.4 installing rgdal and rgeos packages?
Thank you

[s6039642]

This is solved. Thanks to @thho.

[gsw220]

hello，I am a newbie in PSInSAR,so if I can take few
of your time,I hope to communite with you about some questions in this.

[ABraun]

feel free to go ahead.

Just one hint to this post which collects the most important information PS InSAR:

About the StaMPS category

For those who want to get into PSI processing of Sentinel-1 data and prepare it in SNAP, these links were considered most helpful and may provide a good start. New summary on installation, preparation and processing of StaMPS (September 2019) provided by @Gijs Documentation on SNAP-StaMPS workflow Most recent summary summary of the method (February 2019) provided by @m477h145 SNAP - StaMPS Workflow Documentation release of snap2stamps (August 2018): A python-based tool for the automated pr…

[thho]

For all having the same problem, use this as a starting point:

rgdal package installation

[Federica]

Hi Thho, thankyou for all your posts.
I’m trying for the first time to visualize my PS just processed.

I copied the script you suggested ad I got this error

Deramping computed on the fly.
**** z = ax + by+ c
619688 ref PS selected

savename =

'ps_plot_v-do'

Color Range: to 109 m/yr

ans =

[ ]

Undefined function or variable ‘lon2’.

Do you know how to fix this parameter?
thank you very much

[ABraun]

have you executed it line by line, including the plot command beforehand?

So at best you follow this order

ps_plot('v-do', 'ts');

wait for the image to load

load parms.mat;

then execute the rest as a whole.

[wharfbanger]

Greetings StaMPS experts,

I have a question I hope someone can help with. I have processed both ascending and descending Sentinel-1 data for a volcano. The processing worked great as I can clearly identify inflation on part of the volcano for both datasets - a great result. Then I selected only the inflating area (about 300 PS points), and averaged these time series into a single time series for ascending, and another for descending . I notice the descending data provides much better “smoother” average time series curve (below). Can I ask what are the possible reasons for the poor quality/scatter in the ascending time series? Could it be atmospheric effects, or something about that orbit? Is there anyway to analyse the data for an explanation? I would like to explain this.

Thank you kindly,
Mark

[thho]

My first thought on that is a question:

May it be that the radar looks downslope in descending and against the slope in ascending orbit?

If so, in the ascending images the forshortening/layover effects may dominate your signal, hence the poor results…

But this is indeed a good case study to present it, would love to see the studyside in a map and how you images are oriented.

[wharfbanger]

@thho

May it be that the radar looks downslope in descending and against the slope in ascending orbit?

Thanks for your response. The PS selection area is relatively flat (crater floor - see black dash boundary below).

Below is a map showing the descending PS data interpolated by kriging . Both datasets show strong inflation in this area . White area is crater lake.

Please can I ask how to easily show determine the orientation of the satellite look angle? Is this available in the metadata?

[thho]

Concerning my thoughts of forshortening and layover:

• Ascending (south north) looks (always right) downslope the area you use for the plot
• Descending (north south) looks (also always right) “against” the slope (foreshortening/layover)

Hence, Ascending should provide better results, but in your case it performs worse (taken the displacement plot). Why?

Here I am not sure, but my thoughts on that are:

• The slope is not that steep (as you already said) hence the orbit direction (and therefore foreshortening and layover, where the last my not occurre at all) is not crucial in your case, therefore other contributions affects the phase signal
• surely, atmospheric effects could be one point, it would be random, that just the ascending images are affected but it is possible though. Anyway, imagin the case, where the master (since it must be different for both time series) of ascending is highly affected by an atmospheric phase contribution…that could lead to overall more noisy phase signal in the PSI approach, when the atm phase can not be identified correctly during processing.
• What about the amount (number) and distribution (equal or clustered) of PS compared of ascending and descending images…when there are differences in both or one of these aspects, that may be another answer, since your timeline seems to be an average of the area.
• Do you use the same reference point for both? (I assume that, but just to be sure, but anyway, where is it, I am interested and a sign in your map would be helpful )

These are my first thoughts on that, hence PS processing can be very case specific, there are some more details to think about but it is hard to tell, without seeing the interferograms.