Profilometry for the measurement of 3-D object shapes based on regularized filters

Regularized filters (RFs) are used in profilometry for the measurement of three-dimensional object shapes. A linear grating pattern projected onto an object is phase-modulated by the 3-D shape. Phase information of the pattern is obtained by a demodulation process using RFs based on Bayesian estimation theory with Markov Random Fields (MRFs) as prior models. The technique is fully processed in the space domain and applied in the 2-D case to provide a better separation of the height information from noise. As shown herein, the technique is not so sensitive to the singularities in the pattern image as with common filters used in the spatial synchronous detection. The technique and experimental results of real surface profiles are presented. (C) 1999 Published by Elsevier Science B.V. All rights reserved.