flat-earth correction in generate_ifgram.py

[eszterszucshu]

Hi All,

I'm running MiaplPy with a GAMMA-generated SLC stack and I'm trying to understand where the flat-earth (and topographic) phase is removed when interferograms are generated from the phase-linked stack.

Looking at generate_ifgram.py, the interferogram appears to be formed simply as the phase difference between the reference and secondary phase-linked SLCs:

ref_phase = phase_series[ref_ind, ...]
sec_phase = phase_series[sec_ind, ...]

ifg = np.exp(
    1j * np.angle(
        np.exp(1j * ref_phase) *
        np.exp(-1j * sec_phase)
    )
)

I cannot find any step that applies a geometric flattening term (or a simulated topographic phase) before writing filt_fine.int.

My understanding from the ISCE discussions (#513 and #931) is that standard ISCE interferograms are flattened during interferogram formation using the orbit/geometry information, rather than storing flattened SLCs.

Therefore I would like to clarify:

1. Does phase_series.h5 already contain phase-linked SLCs with the flat-earth phase removed?
2. If not, where in the MiaplPy workflow is the flat-earth phase removed before inversion?
3. Is generate_ifgram.py intentionally producing an unflattened interferogram?
4. If MiaplPy is used with the GAMMA workflow, is an additional flattening step expected before unwrapping?

The reason I'm asking is that the interferograms generated by MiaplPy contain a strong deterministic range fringe pattern, whereas the corresponding GAMMA differential interferograms are already flattened. I'm trying to understand whether this difference is expected or whether I'm missing a processing step.

Any clarification would be greatly appreciated.

Thank you!
The miapl ifgram and the one created by gamma
view.py inverted/interferograms_single_reference/20230317_20230329/filt_fine.int --wrap --wrap-range -3.1416 3.1416 --nodisplay

20230317_20230329.diff.bmp

[eszterszucshu]

UPDATE

Could you clarify how the sample GAMMA RSLCs were generated? When I create a raw interferogram using gamma's interf_SLC, the sample dataset already produces an almost flattened-looking interferogram, whereas RSLCs produced by my standard GAMMA workflow retain the expected flat-earth phase. Since MiaplPy simply phase-links the input SLCs and forms interferograms by conjugate multiplication, I'm trying to understand what preprocessing was applied to the sample RSLCs that makes them suitable for this workflow.

I checked that slcStack.h5 is an exact copy of the input SLC, so the difference is not introduced by prep_slc_gamma.py or load_data.py.

Can someone tell me what gamma command was used to create the SLCs? I have release 2020.

Thank you in advane!