Interferogram averaging for DEM generation

Microwave Toolbox Interferometry
103

Hello, i would like to create an average ifgs (Interferogram Stacking) for subsidence mapping. i have some multiple masters-slaves ifg pairs (ab-bc-cd-…)

Can i just use SNAP to complete Step 7 to 10 in the Stack Averaging or i need to use Stamp/Matlab?

Prepare for STACK AVERAGING
STEP 1: master+first slave -> TOPS Corregistration, if necessary Deburst (and possibly subset), interferogram, topographic phase removal,
STEP 2: master+second slave -> coregistration, if necessary deburst (and possibly subset), interferogram, topographic phase removal. STEP 3: master+third slave -> coregistration, if necessary deburst (and possibly subset), interferogram, topographic phase removal. STEP 4: master+fourth slave -> coregistration, if necessary deburst (and possibly subset), interferogram, topographic phase removal.
STEP 5: Create a stack with results of step1,step2, step3, step4. STEP……
STEP 6 Create stacking averaging
STEP 7 SNAPHU EXPORT
STEP 8 UNWRAPPING
STEP 9 SNAPHU IMPORT
STEP 10 PHASE TO DISPLACEMENT

i have tried the averaging using snap but all of the phase are ruined. so, i guess…it is a NO
Later,
i have also come across a post on ID_3102_SAR-EDU_Tutorial_Deformation_monitoring_with_InSAR_pdf about the stacking for Summation of the stacks of Ifgs but no info how to Divide the Sum of Displacement by Total Time

2
2.jpg1214×824 94.3 KB
1
1.jpg1214×824 90.8 KB

i am thinking of doing it in gis/statistically but not sure how

Can anyone provide the steps on Interferogram Stacking? TQ
Dr @falahfakhri may have some info on this?

104

the steps you propose refer to classic DInSAR, and you can do them without StaMPS/Matlab.

The question on how to summarize displacements referring to different periods is challenging, yes. Especially when all slaves refer to the same master, the results cannot be summed up becausee they refer to periods containing each other (AB is within AC, for example).

Maybe someone has a suggestion.

1 Like
105

so, it is easier to sum and divide the small baseline (ab-bc-cd) pairs? since the temp baseline is almost constant

found this image

can anyone suggest the gis steps for it? i believe the Uncertainty,E = Total (ref point - targets)/Ntargets?

106

at least you can add them togethern then and calculate a temporal average displacement, yes.

107

if we want to see the cumulative subsidence during (for example) a year, we should add all displacement bands for each pair using band math? (ab + bc + cd + de) is there another way?

108

I’d say so, yes. An example is given at the end of this tutorial: Sentinel-1 TOPS interferometry

We hope to include pyrate support in SNAP in the near future. It especially targets the multi temporal analysis of DInSAR pairs.

109

thank you so much @ABraun

110

@ABraun regarding the combination of interferometric pairs, what if the different pairs have different Bperp (even values that are considered to small for DEM generation). How does this affects the final average of unwrapped phases. Should these pairs be discarded?

111

Any update on this?
My approach of stacking a phase was to add all the angles together and instead of doing an average, I modulo PI. I think it is more correct at least it looks good and shows some fringe for deformations.
Really interested to know why this cannot go into Snaphu? It seems that this is a serious bottleneck at multiple levels.