Adaptive automatic data analysis in full-field fringe pattern based optical metrology

Fringe pattern processing and analysis is an important task of full-field optical measurement techniques like interferometry, digital holography, structural illumination and moire. In this contribution we present several adaptive automatic data analysis solutions based on the notion of Hilbert-Huang transform for measurand retrieval via fringe pattern phase and amplitude demodulation. The Hilbert-Huang transform consists of 2D empirical mode decomposition algorithm and Hilbert spiral transform analysis. Empirical mode decomposition adaptively dissects a meaningful number of same-scale subimages from the analyzed pattern - it is a data-driven method. Appropriately managing this set of unique subimages results in a very powerful fringe pre-filtering tool. Phase/amplitude demodulation is performed using Hilbert spiral transform aided by the local fringe orientation estimator. We describe several optical measurement techniques for technical and biological objects characterization basing on the especially tailored Hilbert-Huang algorithm modifications for fringe pattern denoising, detrending and amplitude/phase demodulation.