Back Geocoding: Master/Slave Order Different in GPT from GUI

I’ve noticed that the master slave selection is a bit different between the S1TOPS Coregistration tool in the SNAP Desktop and the back-geocoding operator from the gpt command line.

Here are the outputs from running the coregistration from gpt (top) and from the GUI (bottom)

In both cases I used the April 08 SLC as the first product (master right?) but with gpt, the first product became the slave.

Here is the command I ran:

This leads me to the question: Which is correct? To produce a ground-relative deformation map what is the correct master slave selection?
Since a larger distance between the two sensors would produce a positive value for subsidence it would make sense to use the time-2 product as the master, yet the unwrapped phase to displacement operator has a negative value in the expression, so is that already accounted for? 43%20AM

These are the general steps, should be taken in account while selecting the master image,

This procedure has been implemented using GAMMA software,


In SANP, from RADAR tab interferometric and then InSAT stack over view, could be applied to choose the perfect master image according to your data,

In this case , would you please to share your .xml file, to check up it,

1 Like

With only two SLCs, either combination of master slave should have the same absolute baselines. In the GUI, the first product selected is used as the master and Read(2) is used as the slave while in the command line the first product listed is used as the slave and the second as the master. The only difference in swapping the master and slave in this case is a reversal of the sign on the unwrapped interferogram and the displacement map.

I wanted to point out that these two methods produce differing results and so my question is which is the correct order for constraining ground-relative displacement values.

The reality of the movement in your study area shouldn’t be changed, even if they are two passes (swapping between master and slave) the matter is easiest in this scenario case, I already explained this fact in the following posts, I also pointed out to the meaning of sings “+”, “-”

Displacement map

Negative an positive signs

Subsidence and Uplift

Subsidence and Uplift calculations

Thank you for the links, however I am still a bit confused, because the sign is decided by the master slave order combination. Are you saying that choosing the earlier image in a pair as master will produce positive values for uplift and negative values for subsidence? It would be nice if there was a standard convention to follow in the case of these simple two-pass interferograms.

In this document here,, the equation for phase to displacement is defined as 02%20AM
In the snap help documentation, phase to displacement is defined as 13%20AM
So you can see the ambiguity here as one has a negative sign in it and the other doesnt; neither state the frame of reference.

This person found that uplift displayed negative values

However, in the RUS Copernicus training,, they use the first image as master and generate the expected results of subsidence, negative values.

The formula was changed in the past, but as of version 5.0 the phase to displacement operator had been using the negative sign for a while.

No, at all, I dient mean that, the main goal of selecting the master image is, to reference all the slaves geometrically to the master image has been selected , the process plays an important role in this operation is, the corrigestration step, that’s why choose master image has plenty of steps should be followed, as I mentioned in GAMMA S/W case,

Let’s take simple example, the pass A is happened before an earthquake event, and the pass B is acquired after the event,

Selecting A or B as master, doesn’t change the fact of subsidence related to the buildings collapsing,

But the sign is used as + or - referring to the uplif or subsidence is swapping according to the software, as I pointed out in the posts I mentioned to you.