Hard-coded PS coherence in integratePS.py

[stoormgeo]

The final coherence (tcorr_ps_ds.bin) of the PS targets is constantly been 0.95. Then in Mintpy I can exclude PS targets by mask but trying to figure out why PS target integration is needed in the current implementation.
Thx.

[scottstanie]

Hi,
I believe the general idea of the workflow is

• the PS pixels are the most reliable pixels. Their phase shouldn't be altered by multilooking
• After all the other noisier pixels have been processed by the DS processing (meaning the SHP test using nmap, and the multilooking), we insert back the phase from the most reliable pixels where they existed.

I believe the 0.95 was picked as "some very high temporal coherence" in case you wanted to mask out low coherence pixels later. There's no significance to 0.95 exactly, AFAIK.