Efficient de-noising method for Fourier spectrograph

Abstract

The invention relates to the technical field of spectral analysis and more particularly to an efficient de-noising method for a Fourier spectrograph. Compared with the prior art, the present invention provides a Fourier spectroscopic data linear filtering method based on frequency domain analysis. For a sine wave with a single frequency, the peak-to-peak contrast ratio of a target signal subjected to frequency domain filtering is 19.96 which is 12.5% higher than 17.74 that is the peak-to-peak value without the frequency domain filtering, which indicates that the method has better signal restoration performance than a traditional method under the influence of nonlinear noise.

Description

technical field [0001] The invention belongs to the technical field of spectrum analysis, and in particular relates to a high-efficiency denoising method for a Fourier spectrometer, which is a data processing method based on frequency-domain analysis and filtering before spectral diagram reconstruction for the Fourier spectrometer. Background technique [0002] The Fourier spectrometer obtains the spectrogram by Fourier transforming the interferogram, which has unique advantages such as multi-channel, high throughput, high spectral resolution, fast measurement and high signal-to-noise ratio. Standard data processing methods have been formed, including data preprocessing, apodization, phase correction, zero padding and Fourier transform. However, frequency domain filtering is not included as a commonly used noise processing method. The reason is that after the spectrogram is obtained, it can be directly determined according to the characteristics of the hardware, such as the ...

AI Technical Summary

Problems solved by technology

However, frequency domain filtering is not included as a commonly used noise processing method. The reason is that after the spectrogram is obtained, it can be directly determined according to the characteristics of the hardware, such as the wavelength range of the light source, the response range of the detector, etc. This approach is actually based on the assumption that all processing methods before Fourier transform are linear, but this is often not the case, and nonlinear transformation will be introduced in the process of zero padding and phase correction method, so the filtering on the spectrogram is not equivalent to the filtering before the Fourier transform, so in order to remove the noise in the range of non-target spectral bands, it is necessary to carry out a band before the Fourier transform, especially before the nonlinear operation. Filtering to reduce the impact of noise on data accuracy

Images