Automatic selection of frequency domain filter for interference fringe analysis in pulse-train interferometer

The Fourier transform method is typically used to extract the fringe envelope from a pulse-train interference fringe by separating the signal and noise in the frequency domain. Thus, the question of automatic selection of the frequency domain filter to remove undesirable noise naturally arises. When interference fringes are formed, all wavelengths overlap with the same phase at the peak position of the fringe envelope. To realize automatic selection of the frequency domain filter, we observe the changes in the peak position point of the interference-fringe envelope and the phase-crossing point of different wavelength components in response to variations in the filter position and size. Hence, we select the frequency domain filter that yields the minimum difference between these two points. The proposed algorithm is verified using experimental data. This technique can be applied not only to white-light surface profilometry, but also to pulse-train length metrology.