Machine Vision Assisted Precision Alignment System for Optical Path Calibration in Point Diffraction Interferometer

In point diffraction interferometer (PDI), alignment error between objective convergent spot and diffraction pinhole is inevitable during the tests. This phenomenon will affect the performance of PDI in follow aspects: 1) great wavefront error; 2) reduction on diffraction efficiency; 3) quality of interferograms. All these factors may finally reduce the measurement precision of the instrument. Aimed at these technical challenges, we propose machine vision assisted precision alignment system for PDI optical path calibration. Firstly, Rayleigh-Sommerfeld vector diffraction theory is used to build mathematical model among alignment error, diffraction wavefront error, numerical aperture, and pinhole size. To satisfy the requirement of error calibration, scheme of machine vision assisted optical path alignment system is designed. Then we adopt magnetron sputtering and electron beam lithography to fabric precision diffraction pinhole. In this stage, alignment images as well as PDI diffraction efficiency (intensities ratio between reflected and diffracted beam) are the dominant information to determine alignment error in multi-directions. In addition, specialized image processing algorithm is designed that can measure alignment error in pixel and physical scope. To realize automatic correction, mathematical model between measurement and control quantities is constructed. Finally, implementation and experiment of this method are also introduced. Misalignment on lateral translation (XOY plane), longitudinal defocus (optical axis Z) and tilt error (XY tile) are well calibrated and the quality of interferograms is also improved. It can be concluded that the proposed system has advantages in accuracy and efficiency of optical path alignment.