Analysis and suppression of random noise effect in interferometry on absolute measurement of shift rotation

A statistical model was constructed to study the effect of random noise in interferometry on the absolute measurement of shift rotation. The results indicated sensitivity to random noise, particularly under low signal-to-noise ratio conditions. Therefore, we propose a low-signal-to-noise-ratio surface reconstruction (LSSR) algorithm, which is a compelling iteration for achieving the surface reconstruction of absolute measurement of shift rotation. In simulations, the peak-to-valley (PV) of the residuals of LSSR reached λ/1000, whereas the classical method could only reach λ/100. In experiments, the PV of the surface distributions reconstructed using LSSR reached λ/40, and the reconstructed surface distributions were consistent regardless of the magnitude of random noise. In addition, compared to the three-flat mutual inspection method, LSSR also exhibited consistency, significantly suppressing the effect of random noise on the absolute measurement of shift rotation.