StaMPS-Visualizer, SNAP-StaMPS Workflow

For your understanding: the LOS displacements are relative to each other and to their flight direction, so its important to set a reference point and combine ascending and descending values to make the values absolute.

After the reference point you can use the formula: LOS displacement / (cos(radians(incidence_angle)) to convert the LOS displacement to vertical displacement. When you have this vertical displacement you can combine the ascending and descending displacement by the mean of the points on the same coordinates, if you dont have both ascending amd descending on the same point you have to drop this point.

You can use the velocity from the csv files from the stamps-visualizer manual to make this all work. I will add the example scripts later when i am on my laptop.

Specifically for your case you can use the coordinates of the gps station in your aoi as reference point.

4 Likes

Had you seen the snap2stamps scripts? I guess they do similar things (maybe less) than the ones you explained.

Normally points do not match (until the last decimal on the coordinates, as for example StaMPS does not correct the point location, just get their position from the radar coordinates. In my opinion, you better search within a radius for PS points near enough to consider them as having equal or similar characteristics regarding deformation pattern. I hope this helps.

1 Like

@mdelgado

Yes i used snap2stamps and its indeed my idea to create similar scripts but then for this task, scripts are done but will look in hoe to make them pythin wrappers looking at the snap2stamps scripts later.

I rasterize the ps points to the merge_resample_size grid and if the raster points overlap they get combined. For some cases the buffer approach may be more viable but this is dependant on your spatial variability for sure. Generalizing your data for sure isnt always the best approach and “interpolating” like this creates false spatial autocorrelation aswell. But either kriging interpolating, buffer or strict approach is a parameter i intend to implement. I would like to hear more about everyones ideas for this!

Glad to hear that you use my scripts. Please refer or acknowledge them somehow when possible. I will be happy to read about it.

Regarding searching on radius or rasterizing points, I do not see much differences, as you give an area of action to a PS point, which can be sometimes consider pretty much similar. Indeed the radius approach depends a lot on your PS density, and the radius should vary from case to case.

Discussion is welcome and I will be happy to read again about it.
Cheers

Hope it can get published but i am not quite sure of that, the paper with you and foumelis is already in my reference list so that is sorted :slight_smile:

Cheers

1 Like

We are about to release a newer version to include multi-subswath support for TOPSAR and full workflow also for stripmap sensors.

I hope that this is ok for you and not overlapping with your work

3 Likes

With multi-subswath do you also mean combining 2 path/frame combinations?
This is the only thing i hoped was available for SNAP in general, I had no other holdbacks in my processing with snap2stamps although i could imagine a plot function for a quick look at the ifgs or a png converter could be very handy.

Do you mean to combine to slices from same frame? Combining 2 path/frame you can only do after StaMPS,not using SNAP, when your scope is the phase information (deformation or so)

Ouh! Thanks for the reminder… creating quicklooks for ifg should be easy to integrate.

1 Like

while working in Rstudio it exits autiomatically while installing the libraries… Could anyone suggest me a solution to this issue…

i get such an error… any way to solve this?

looks like the data you read into are characters not numerics…can can you provide a screenshot of all four windows of Rstudio…I would like to have a look on the code (upperleft) from line 1 until as much as possible, your last actions before the error in the console (lower left) the environment (upper right) with the objects loc and I assume dat

in the end, I think yomething is wrong with your csv file but that is a guess until now

and is the installing packages thing still a problem? For a first try you can use the gui instead of the script…in the lower right window there are some tabs, one is called packages…you can isntall them via this menue…should be self explanatory…if not, tell me where you stuck

for the installing thing at times rstudio gets crashed and closes… now no issue with it… i also tried installing using gui…


this is the screenshot of my screen

ok as you can see in the Environment (upper right) the objects lat lon and therefore loc are empty…please provide the output from:

colnames(pnts)

or easier, provide a screenshot of your pnts tab in the upper left panel

for some reason, there are points in the lines:

lon <- pnts$export_res_.1[2:nrow(pnts)]

and lat (line beneath)

I am talking about the point after the underscore before the 1 or 2 respectively. delete them…I don’t know what they are doing there. should work after that.

and
in the line 30 where you write the csv file, you write to the same file name as your input file…you might not want to overwrite it…maybe you should add something like “…_subset” to the csv filename in line 30 before overwriting your input data

i did as you said… again a new error

sorry i sorted the mistake
and it went through well

1 Like

easy, you added col to line 22 before pnts.geo by accident in the past…probably when you tried to call clonames() function or something…delete the col , run line22 again and you are fine

1 Like

yup… thank you lots @thho

1 Like