Adaptive double sinusoidal-assisted EMD 3D shape profilometry measurement method

To enhance the three-dimensional (3D) reconstruction accuracy using the Fourier transform profilometry (FTP) method in non-uniform light intensity environments, an adaptive double sinusoidal-assisted empirical mode decomposition (ADSAEMD) algorithm is proposed in this paper. The background is pre-processed employing an adaptive filter, and the fringe density map is used to initially filter out the background of the deformed fringes. Adaptive double sinusoidal signals are symmetrically added and phase-shifted as auxiliary signals, the empirical mode decomposition (EMD) algorithm is used for image mode decomposition. The intrinsic modal function (IMF) generated by iterating through both internal and external layers effectively avoids the mode mixing problem (MMP), and the stopping criterion is set to terminate the iterative procedure of EMD. Simulation and experiment show that the proposed algorithm in this paper separates random noise and background components from the fringe pattern in the non-uniform illumination environment, removes the influence of non-uniform ambient light, and effectively recovers the 3D shape of the object.