Fourier Transform Profilometry Based on Adaptive Gaussian Filter

Fourier transform profilometry (FTP) is an active optical 3D measurement technique that reconstructs object shapes by analyzing deformed fringe patterns modulated by surface height. This paper proposes a binary fringe projection Fourier transform profilometry algorithm employing an adaptive Gaussian bandpass filter. Binary fringes can effectively mitigate phase errors caused by the Gamma nonlinearity of projectors. The adaptive Gaussian bandpass filter automatically calculates the center frequencies and bandwidths of each harmonic component based on the captured binary fringe images. With only two captured grating images, the algorithm combines Fourier transform profilometry with a multi-frequency grating phase unwrapping method to achieve the functionality of multi-frequency grating Fourier transform profilometry using four sets of different frequencies, yielding a root mean square error of 0.028 mm. Finally, the feasibility of the proposed method is verified by comparing its measurement results with those of other multi-frequency grating Fourier transform profilometry techniques.