Phase reconstruction and unwrapping from holographic interferograms of partially absorbent phase objects

A method for automated phase reconstruction from holographic interferograms of nonideal phase objects based on a two-dimensional Fourier transform is described. In particular, the problem of phase unwrapping is solved because earlier techniques are inappropriate for the phase unwrapping from interferograms of partially absorbent objects. A noise-level-dependent criterion for the binary mask that defines the unwrapping path for the flood algorithm is derived. The method shows high noise immunity, and the result is reliable provided that the true phase is free of discontinuities. The phase distribution in the outmasked regions is estimated by a linear least-squares fit to the surrounding unwrapped pixels.