TopSAR Processing


same reason


Maybe my images are not sutable!!:disappointed_relieved:


How can I know? I don’t know your images or your PC environment? I can only guess from the very little information you provide. You didn’t even tell which sensor you are using.


I am using sentinel-1 images for analysing land subsidence in two regions of Iran, before this I was working on another region…and everything went well and without any error like this! But in this new region I just perform processing on 2 monthes data! After that I got into this problem,like java heap script or GC overhead and etc…:disappointed:


did you reduce your data size in the TOPS coregistration menu?
It is described here: The Order of DEM Creating Steps


I didn’t reduce number of bursts before, but now how can I khonw how many number of burstes arr neccesary for me!I set the number between 4 to 6! and my intersted region locates on IW3…
Then Is it right in order to work on IW3 ,we select the number of bursts between 4to 6.???


you can coarsely guess it a bit from the kmz file in the ‘preview’ folder


Thanks alot!!:blush:
I,ll try it​:pray::pray::pray:


I have a very big problem.
I hve followed all the steps:

  1. S1 TOPS coregistration with ESD
  2. Deburst and Subset
  3. Interferogram Formation
  4. Topographic Phase Removal
  5. Goldstein Phase Filtering
  6. Snaphu Exprort
  7. Unwrap via Snaphu
  8. Snaphu import
  9. Phase to Displacement
  10. Range - Doppler Terrain Correction
  11. Choose a reference point
  12. Final map
    This is a basic processing

and I have also create the following band math:
final terrain corrected defo band * coherence mask band.

I have done this process in an area where I’m sure that there is a vertical displacement of -10m but my results are completely different.
How is it possible?
Where can I be wrong?

Please help me.


What are the results? maybe you don’t see any displacement because your area of interest is not coherent?


No, I have done the same process also not considering the band math with the coherence but the results do not coincide.
There should be a displacement of -10m and I have found -1mm and I cannot understand why.
In this moment I have a doubt that maybe is another issue:

when I consider only pixel with high coherence (for example >0.4) I will delet the pixel with a displacement or not?
I have understood that high coherence means that the pixel of the master and the pixel of the slave are equal.
Is it right?
So if I want to obtain only different resuls I have to choose low coherence and not high coherence.
In any case the big problem remains.
Could you help me?


so why don’t you use the area with -10 meters as the reference point and adjust the raster accordingly?


Thanks for your reply.
If I proceed like you suggest me maybe in this case the results are correct but in other case I don’t know what is the displacement so I cannot apply the same method.
I have tried in this area only in order to understand if my job works but like I said before there is a big difference between my results and the effective displacement.
I have understood that the better method is the PS but I thought that the method that I have applied was good in order to obtain vertical displacement


What do you mean by 10m displacement? In what time-frame? Between the SAR acquisition dates? Is this a landslide, ground-subsidence or a tectonic movement? (probably not the last one due the size).


High coherence means that the scatterers in that area are not disturbed.


Si it means that there is no displacement in these points?
If so, if I want to get only the movements, I have to choose low coherences?


No, it means that if there was displacement the whole area moved more or less as a block so that the scatterers were not disturbed. Low coherence areas are just noise and are not useful. Please give more info about the type of displacement (see my other question).


I’m sorry but I had not seen your question.
I’m working on a new project where we have to found the vertical displacement of the surface.
I have looked for info and manual on the web and I have followed all the steps found on the user manual of the Interferometry tutorial of ESA that coincide with what I have found on this forum.
So, I have used SNAP and SNAPHU like ESA suggests.
When I have finished to processa an area, I was not sure that my results were correct because I have not a feedback so I have repeated the same process in an area where I know that a street dropped down in Rome for -10m.
So I have used two S1_SLC images where the master was before the fact and the slave, after.
But, as I said before, the results are completly different.


That disturbs the scatterers completely, plus the movement is so large that it would anyway be impossible to tell how many multiples of the wavelength the surface moved. So this is not going to work - you could try to detect precursors of the collapse (if any) by doing PSI-processing with SNAP+StaMPS.


Yes thanks.
I have already been advised to use Stamps.
My problem is that it was months that I was trying to get this result and now I realize that I have to use another very complex method that will make me lose a lot of time that I do not have.
Also because others in the forum have sent me some documentation to study but very theoretical, but I have not found any guide telling me step by step what to do to get the vertical dislacement with snap and stamps.
I would need one like the one I found ESA in which it explains in detail the steps to follow on snap, how to install snaphu and how to get out snaphu.
Can you help me on this?
I also asked if there are courses on the subject but at the moment there are no online and the first profit is in September.
I would need a manual like the one I said or someone to give me lessons on how to proceed.