Evaluation of single-shot and two-shot fringe pattern phase demodulation algorithms aided by the Hilbert-Huang transform

Abstract

In this contribution we evaluate single and two-shot techniques, namely the Hilbert spiral transform (HST) and the Gram-Schmidt orthonormalization (GSO) in terms of phase demodulation accuracy in the complex fringe patterns analysis (i.e., with strong background/contrast variations, severe noise, considerable local gradients of fringe shape/orientation). Both methods are aided by the novel Hilbert-Huang transform (HHT) processing to adaptively reduce demodulation errors. The HST utilizes a spiral phase function and a spatial fringe orientation map to demodulate phase of complex fringes. It is especially susceptible to uneven bias term and noise. The HHT method realizes bias/noise suppression adaptively with outstanding accuracy. The GSO is a fast two-shot fringe-shape-robust phase demodulation scheme. It treats two arbitrarily phase shifted interferograms as vectors and conducts orthogonal projection of one vector onto another. The GSO is susceptible to background, contrast and noise fluctuations, however. The HHT method is perfectly suitable to perform efficient pre-filtering. Both methods (HHT-HST and HHT-GSO) are proven versatile and robust to fringe pattern defects using simulation and experiment.