* standard: usual standard unwrapping method, pixel by pixel, it suffers the usual drawbacks of unwrapping in presence of noise. If you mask an interferogram and you have disconnected areas, it’s not going to connect them.
* sparse LS: it uses sparse points and Least Squares to connect unconnected areas and to solve the problem in presence of noise. It’s by construction smooth, good if you have no alternatives to return a smooth result
* other options: prototypes and experiments, don’t us them…