How to prepare Sentinel-1 images stack for PSI/SBAS in SNAP 5


#123

Glad to hear that you solve your problems.

Now remember that due to the bug on topo phase removal the results may be wrong. Hope that they fix this bug soon.

Joaquin


#124

Dear Sharon,

       Thank you very much relly!I will try what you said.Thank you.

#125

By cheking some control points in the borders and comparing their coordinates before and after terrain correction some shifts might appear, like in my case…


#126

Dear @Sharon ,

Have a look one this poste :

deburst must be performed after ifg formation.

Good luck.


#127

Dear all,
First of all thanks for sharing your work and findings (@FeiLiu, @katherine, @annamaria…), it saved me a lot of time of research and testing!
I tried to perform the PSI analysis with Stamps and I’ve encountered the same problems with the topo phase removal operator. Then I tried the stack approach explained by @annamaria and the results are much better. However, it seems that there is a problem with the metadata when performing the interferogram stack using the Create Stack tool and Stamps is unable to retreive the baseline information from .base files in diff0 folder (after using the stamps export tool). Has anyone faced the same problem?

Albert Vega Valeri


#128

@alvegavaleri i don’t undestand what is your problem.
By change in your .base files appears NaN values? Or there are problems during StaMPS elaboration?


#129

@annamaria Exactly, in some ocasions NaN values appear in the .base files. I don’t know which is the problem because it only happens sometimes… Do you know any easy method of retrieving this data manually and modify the .base flies?


#130

Hello Katherine,

First of all, thanks for sharing this post about staMPS PIS. It seems you helped many people to perform PSI technique.

I have a question.
As we know, in order to properly do InSAR, DiINSAR nad PSI techniques there are some strict rules for deriving optimum results. Two of these rules are the following:
-Satellite orbital path must be the same for both satellites
-Orbit direction must be the same (both ascending or descending orbits)

We may have 15 Sentinel-1A/B images on the archive for one month but only 2 or 3 of them have the same orbital path and orbit direction. In case we need to collect 30 (roughly) S1 images during a 6 month period to perform PSI, we cannot stick to the rules as it is impossible to find so many S1 images with the same orbital path and same orbit direction. So, we might need to use both ascending and descending orbit, or S1 image acquired from different orbits. If we do something like, do you think that this is going to affect a lot our results?

According to your post, how many S1 images did you use? were they all acquired in the same orbital path and direction?

Thank you
Ioannis


#131

@alvegavaleri
In my experience, this problem appears (only) when coregistrated images and ifg hasn’t the same name.

  1. I created a stack of these 4 results / pairs (with Create stack tool). In this stack, the images must be renamed with a name compatible with what appears in layer1 (for example, if the pair is 29Apr2015 and 22Jul2015, in layer1 you have i_IW1_VV_mst_29Apr2015, q_IW1_VV_mst_29Apr2015, i_IW1_VV_slv_22Jul2015, q_IW1_VV_slv_22Jul2015, in layer3 your renamed image must be i_ifg_VV_29Apr2015_22Jul2015 and q_ifg_VV_29Apr2015_22Jul2015).

Let me know.


#132

Yeah, I saw your post and I renamed the ifgs but in some occasions the problem persists… I’ll let you know if I found something. Thank you very much!


#133

Hello Ioannis:

To perform a dinsar analysis all scenes must be from the same track, which implies same relative orbit and acquisition geometry. You cannot mix images from different tracks because the perpendicular baselines will be bigger than critical baseline and the acquisition geometry will cause an useless interferogram.

For PSI a minimum of 30 images is recommended in several papers, but you can get results with a smaller dataset. Also if your zone doesn’t have good temporal coverage you can increase the period over 6 months, there is no time limitations there.

Joaquin


#134

@krasny2k5 thank you for your answer.

I understand that for PSI technique, we need many images. So, let’s say we need to perform PSI to detect a subsidence in an urban area for time interval of 5-6 months and we do not have this amount of Sentinel-1 images required (25-30) within this period of time, would it be silly to combine images from different orbit directions (ascending and descending) in order to reach 25 images? or this will ruin our interferograms?

thanks
Ioannis


#135

No, you can’t generate an interferogram between ascending and descending acquisitions.


#136

Dear alvegavaleri,

I met the same problem as you do, sometimes the output .base file appears NaN value, so I just redo the export step until I get the right value.


#137

yes, I understand.
thanks


#138

Hi Joaquin,
Many thanks for your help.
I have a doubt regarding the deburst process of the interferogram. I have read both, that the deburts process have to apply BEFORE and AFTER the interferogram formation.
What is the right answer?

Thanks again.


#139

Hi @AranLarra

You should deburst after coregistration process and before interferogram formation.

Joaquin


#140

Hi @krasny2k5

Thanks for your fast answer.

Another doubt that I have in mind… I am searching ground deformations using the PS technique, soall the interferograms are formed using the same master image as reference? That is to say, if I have an image serie from january 2015 to january 2017 and the master image is from march 2016 for example, all the interferograms are formed respect to the image of march 2016?

And a last question, what about the polarization? Is there any interesting polarization? In the case that the image has two polarizations, VV and VH, which one should I use?

Many many thanks in advance.

Aran.


#141

all the interferograms are formed respect to the image of march 2016?

Yes.

In the case that the image has two polarizations, VV and VH, which one should I use?

Usually VV polarization is used in DinSAR tecnique.


#142

Thak you very much @annamaria,
All your answers are of great help for me.