A near infrared spectrum signal denoising method based on a Hodard-Presscott filter

Abstract

The invention discloses a near infrared spectrum signal denoising method based on a Hodard-Presscott filter. The method comprises the steps of constructing a decomposition formula of the spectral signal according to the Hodard-Presscott filter, wherein the input value of the decomposition formula comprises a regularization parameter and a near-infrared spectral signal dyed with noise, and the output value of the decomposition formula is a near-infrared spectral signal not dyed with noise. Compared with the existing near infrared spectrum denoising, the method has a better denoising effect andcan provide a better basis for analysis and detection of fruit and vegetable pesticide residues.

Description

technical field [0001] The invention belongs to the technical field of near-infrared spectrum signal processing, and in particular relates to a method for denoising near-infrared spectrum signals based on a Hodrick-Prescott filter. Background technique [0002] Near-infrared spectroscopy refers to the near-infrared band in the range of 780nm-2500nm created by the absorption spectrum, which can effectively reflect the characteristic information of hydrogen-containing groups (C-H, O-H, N-H, etc.) in organic molecules in samples. Based on this, near-infrared spectroscopy Analytical technology can be used for qualitative, quantitative analysis and detection of compounds containing C-H, O-H, N-H and other hydrogen groups. It has the advantages of no damage, no pollution, good analysis repeatability, simplicity and speed, and is widely used in the field of food testing. . [0003] However, the spectral signals measured by the near-infrared spectrometer usually contain both useful...

AI Technical Summary

Problems solved by technology

This leads to the situation that it is difficult to choose the number of sampling times: the number of sampling times in the actual measurement is too small, the spectrum is greatly affected by the randomness of the sampling, and the spectral accuracy is low; cannot be removed by the method of taking the average value) and increases accordingly, which not only sacrifices time, but also cannot guarantee spectral accuracy

In recent years, the wavelet threshold denoising method (such as Chinese invention patent CN104020136B) has been used to denoise near-infrared spectral signals, but this method is prone to Gibbs oscillations in discontinuous point areas, causing spectral signal distortion and destroying the geometry of the spectrum. characteristics, affecting the analysis and identification of chemical groups

[0005] In the 1920s, Whittaker proposed the Hodrick-Prescott decomposition method. Since the 1990s, it has been widely used to separate economic trends and shock waves at various time points in economic data in economics. This method is rarely used in the field of near-infrared spectral signal processing. Research

Images