GAMMA Questions

Hello Kang,

Thank you very much for your question. The origin of the missing unwrapped
points has been interesting. There is now a simple solution that has been
implemented and tested.

Your example illustrated an oversight by mcf_pt in the actual phase unwrapping
that is performed after the minimum cost flow problem has been solved.

The MCF program creates a triangular mesh of the valid pixels using Delauney
triangulation. The phase inconsistencies (residues) are calculated for every
triangle in the mesh with a positive residue +1 and a negative residue -1.
Network flow of the residue charge is optimized between nodes of network
constructed from the triangles using MCF optimization.

Unwrapping proceeds by "growing" the phase starting at a corner of a seed
triangle and looking at adjacent neighbor triangles. If an adjacent triangle has
points that have not yet been unwrapped, that triangle is added to a second
list. We ping/pong between triangle lists filling the entire connected region
while taking into account the integer MCF network flow solution. If flow is
present, then there is an increment of multiples of 2PI added to the phase of
the points in the adjacent triangle.

In some special geometrical situations, it is possible that a triangle can be
missed, and a single point of that triangle is not unwrapped. This omission has
been corrected so that we now scan through the image at end of the ping/pong
triangle unwrapping step and identify any points that were missed, and unwrap
the triangle containing that point. In the example you sent us, there were only
4 points that were not unwrapped out of 100000.

For which OS do you want a correct version of mcf and mcf_pt?
What was the version of the Gamma software that you downloaded and installed?

Best regards and thank you for your careful observation!

Charles and Christophe

On 23.08.2024 23:33, Kang Liang wrote:

GAMMA Questions - IPTA / SBAS: `mcf_pt` mask some output pixels <https://
dep1doc.gfz-potsdam.de/boards/25/topics/661>

/Kang Liang/

Dear GAMMA developer,

I find the several 0.0 value pixels in the unwrapped data generated from
`mcf_pt` while no pixels in the input data are masked or 0.0.

Here is an minimum example:

```
mcf_pt plist - pint - - - punw
```
The `plist` and `pint` are attached here (note that I manually add the .png
extension to meet the requirement of this forum) and there are no masked pixels.
But the generated `punw` has 0.0 value pixels.

I am using GAMMA20231208.

Best,
Kang

Files pint.png <https://dep1doc.gfz-potsdam.de/attachments/download/1001/
pint.png> (781 KB)
plist.png <https://dep1doc.gfz-potsdam.de/attachments/download/1002/plist.png>
(781 KB)

---

You have received this notification because you have either subscribed to it, or
are involved in it.
To change your notification preferences, please click here: http://dep1doc.gfz-
potsdam.de <http://dep1doc.gfz-potsdam.de>