采用小波阈值的时变光栅莫尔信号去噪方法

Denoising method of dynamic grating Moiré signal based on wavelet threshold

莫尔光栅的纳米级测量需要对莫尔信号进行高倍细分,而高倍细分的精度往往受到高斯白噪声的影响。将莫尔信号视为稳态模型进行去噪分析与处理时存在信号频率相对固定的缺陷,根据信号的频率是大范围可变的,且噪声分布在整个频率范围内,提出了一种更符合实际的时变模型,并采用小波阈值去噪法对信号进行处理。对时变莫尔信号进行了建模,对小波去噪原理及阈值去噪法进行了分析,经大量实验对比,选用Sym8小波基、分解尺度为6、阈值准则为Heursure的软阈值法去噪效果最好。去噪后,光栅莫尔信号接近理想信号,使莫尔信号的细分倍数达到1000倍。

Nanoscale measurement of Moiré grating requires the high subdivision of Moiré signal. But the precision of high subdivision is often affected by Gauss white noise. There exists many deficiencies that the frequency of signal is relatively fixed when Moiré signal is considered as a steady model. A factual dynamic model was proposesd, in which the frequency of signal varied in a wide range and the white noise usually distributed in the whole signal frequency band. And then a wavelet threshold denoising method was adopted to process the Moiré signal. Firstly, the dynamic model was established and the principle of wavelet denoising and the methods of threshold denoising were analyzed. Every parameter was selected through the vast simulation experiments. The experimental results show that the denoising results of Moiré signal are best when adopting Sym8 wavelet basis, the ＂Heursure＂ rules of soft threshold selection and the scale decomposition of 6. After denoising, the Moiré signal is near to the ideal one, which makes the signal′s subdivision reach 1 000 times.