大口径平面镜子孔径拼接检测中表面中高频误差的检测误差处理方法研究

The testing errors of mid-high frequency surface error in large optical flat mirror sub-aperture stitching testing and errors processing methods will be studied in this project. Based on the experimental data of sub-aperture stitching testing of flat mirror, the effect on the SlopeRMS testing noise of different number of averaged data will be analyzed, then the number of averaging data can be determined. Several data filter methods and pixel fusion method will be applied to the noise data, and the best method for noise reduction will be chosen by the comparison results. To realize the calibration of system mid-high frequency error such as standard reference surface in the sub-aperture stitching testing, an algorithm employing sub-aperture stitching and noise reducing method will be explored, with optimization of slope noise data. Stimulation will performed to examine the algorithm, and the accuracy will be analyzed. Finally, the sub-aperture stitching experiments will be arranged and the new algorithm will be applied to the experimental data, the Instrument Transfer Function will be involved to analyze the accuracy of the sub-aperture stitching system. The results will be used to compare the algorithm with the small aperture averaging method. Several experiments will be done to get the repeating accuracy of SlopeRMS. The research will improve the accuracy of mid-high frequency error in sub-aperture stitching testing, so effective testing data can be obtained to ensure the fabrication of large flat mirror.

本项目针对大口径平面镜子孔径拼接干涉检测中尚待解决的中高频误差的检测误差处理方法问题开展研究。首先根据子孔径拼接检测平面实验数据，分析不同数据平均次数对SlopeRMS检测噪声的影响，研究确定合理的数据平均次数。通过对比分析不同滤波方法和像素融合法对检测噪声的抑制效果，确定最优的降低噪声方法。其次，研究以斜率噪声最小为优化目标的子孔径拼接与降噪声方法相结合的系统误差分离算法，抑制噪声对算法的影响，实现中高频系统误差的有效分离。运用仿真手段验证算法，并分析算法精度。最后，设计子孔径拼接实验并应用系统误差分离算法，结合干涉仪的仪器传递函数，综合分析经过误差处理后的检测精度，并与小孔径多次平均绝对检测方法的结果进行精度对比；开展多次检测实验，获取误差处理后SlopeRMS的重复性测量精度。本项目通过提高子孔径拼接检测中的中高频误差检测精度，为大口径平面镜的制造提供高精度检测数据保障。

本课题通过拼接检测中子孔径数据的重叠相关性，结合非线性误差解耦分离手段，对子孔径拼接检测中的电子学噪声、气流、温度等环境干扰进行了定量评价、噪声抑制、算法设计和实验验证。通过检测噪声抑制方法的对比研究，优选出像素平均法作为降低Slope-RMS检测噪声的最佳方法，实现了单点0.21μrad的Slope-RMS检测精度。开发了一套针对大口径平面镜子孔径拼接自标定软件，仿真实验表明，拼接检测精度高达1.1nm/R Slope-RMS。本项目中所使用的理论分析、数学建模、光学仿真、模拟检测和实验方法的总结和成果梳理工作正在进行。综合上述结果，本研究为探讨面形检测中高频误差的抑制机制奠定了坚实的工作基础。同时对于认识子孔径拼接与误差分离的关系提供了重要的启示性线索。