Correlation coefficient algorithm for high-precision white-light interferometry

We propose a filtering algorithm that ombines white-light interference signal (WLIS) characteristics and the Pearson correlation coefficient. Unlike the widely used WLIS preprocessing algorithm (fast Fourier transform, phase-shift algorithm), the proposed algorithm takes advantage of the fact that WLIS is periodic and accurately distinguishes between useful and noisy signals. For a proofed imaging system (center wavelength 635 nm, 10× Mirau interferometric objectives, and step size 24 nm), the proposed algorithm recovers a smoother surface compared with the algorithm mentioned above. Moreover, the proposed algorithm has a better processing effect and higher timeliness than the traditional WLIS processing algorithm. Compared with the conventional preprocessing algorithm, the root mean square error of the recovered surface is reduced by about 2 times, and the time consumed is reduced by 1.5 times on average. It is believed that we provide a practical method for high-precision measurement.