I want to know that what’s the method of SARPROZ choose master image? Can you tell me that? Does it based on temporal baseline or spatial baseline，or other method？ Is there has some literature about how SARPROZ choose master image？
And SARPROZ use which method on image Registration？ Is there has some literature about SARPROZ‘s method？
The master is chosen in the barycenter of the distribution of normal/temporal baselines.
If you are mixing multiple polarizations, the sw will take a master among images at the highest resolution.
Registration is carried out by cross-correlating patches of images.
Then a polynomial model is fitted in the estimated shifts considering their SNR.
Goodness of fit is used to keep points and discard outliers.
Only after precise geocoding, you have the option to use a DEM to modulate the final registration.
You can find some descriptions in the manual, nothing more has been published (techniques are well known by people working in the field).