Performance evaluation of two-dimensional phase unwrapping algorithms

We present a performance evaluation of eight two-dimensional phase unwrappi ng methods with respect to correct phase unwrapping and execution times. Th e evaluated methods are block least squares (BLS), adaptive integration (AI ), quality guided path following (QUAL), mask cut (MCUT), multigrid (MGRID) , preconditioned conjugate gradient (PCG), Flynn's (FLYNN), and Liang's (LI ANG). This set included integration- (path following), least-squares-, L-1- , and model-based methods. The methods were tested on several synthetic ima ges, on two magnetic resonance images, and on two interferometry images. Th e synthetic images were designed to demonstrate different aspects of the ph ase unwrapping problem. To test the noise robustness of the methods, indepe ndent noise was added to the synthetic images to yield different signal-to- noise ratios. Each experiment was performed 50 times with different noise r ealizations to test the stability of the methods. The results of the experi ments showed that the congruent minimum L-1 norm FLYNN method was best over all and the most noise robust of the methods, but it was also one of the sl owest methods. The integration-based QUAL method was the only method that c orrectly unwrapped the two interferometry images. The least-squares-based m ethods (MGRID, PCG) gave worse results on average than did the integration- (or path following) based methods (BLS, Al, QUAL, MCUT) and were also slow er. The model-based LIANG method was sensitive to noise and resulted in lar ge errors for the magnetic resonance images and the interferometry images. In conclusion, for a particular application there is a trade-off between th e quality of the unwrapping and the execution time when we attempt to selec t the most appropriate method. (C) 1999 Optical Society of America. OCIS co des: 100.5070, 120.3180.