ZHUANG Xin-gang1,2,3, SHI Xue-shun1,2,3, LIU Hong-bo1,2, LIU Chang-ming1,2,
ZHANG Peng-ju1,2, WANG Heng-fei1,2
（1. The 41st Research Institute of China Electronics Technology Group Corporation, Qingdao 266555, China;2. National Opto-Electronic Primary Metrology Laboratory, Qingdao 266555, China;3. Science and Technology on Electronic Test & Measurement Laboratory, Qingdao 266555, China）
Abstract： As one of the important indicators of spectrometer, signal-to-noise ratio (SNR) reflects the ability of spectrometer to detect weak signals. To investigate the influence of SNR on the prediction accuracy of spectral analysis, we first introduce the major factors affecting the spectral SNR. Taking green tea as an example, the influence of spectral SNR on the prediction accuracy of the origin identification model is analyzed by experiments. At the same time, the relationship between the spectral SNR and prediction accuracy of spectral analysis model is fitted. Based on this, the common methods for improving the spectral SNR are discussed. The results show that the accuracy of the prediction set model first decreases slowly, then decreases linearly, and finally tends to be flat as the spectral SNR decreases. Through calculation, in order to achieve the prediction accuracy of prediction model reaching 90% and 85%, the spectral SNR is required to be higher than 23.42 dB and 21.16 dB, respectively. The overall results provide certain parameters support for the development of new online analytical spectroscopic instruments, especially for the technical indicators of SNR.
Key words： near infrared spectroscopy; signal-to-noise ratio (SNR); partial least squares (PLS); spectral analysis; green tea
CLD number： O433.4doi： 10.3969/j.issn.1674-8042.2020.03.002
References
［1］Bgel T, Osinenko P, Herlitzius T. Assessment of soil roughness after tillage using spectral analysis. Soil and Tillage Research, 2016, 159： 73-82.
［2］Landau S, Glasser T, Dvash L. Monitoring nutrition in small ruminants with the aid of near infrared reflectance spectroscopy (NIRS) technology： A review. Small Ruminant Research, 2006, 61(1)： 1-11.
［3］Zhuang X G, Shi X S, Wang H F, et al. Rapid determination of green tea origins by near-infrared spectroscopy and multi-wavelength statistical discriminant analysis. Journal of Applied Spectroscopy, 2019, 86(1)： 76-82.
［4］Wolf M, Quaresima F V. Progress of near-infrared spectroscopy and topography for brain and muscle clinical applications. Journal of Biomedical Optics, 2007, 12(6)： 062104.
［5］Moses W J, Bowles J H, Lucke R L, et al. Impact of signal-to-noise ratio in a hyperspectral sensor on the accuracy of biophysical parameter estimation in case II waters. Optics Express, 2012, 20(4)： 4309-4330.
［6］Banas K, Banas A M, Heussler S P, et al. Influence of spectral resolution, spectral range and signal-to-noise ratio of Fourier transform infra-red spectra on identification of high explosive substances. Spectrochimica Acta Part A： Molecular and Biomolecular Spectroscopy, 2018, 188： 106-112.
［7］Li G, Zhao Z, Wang H Q, et al. The effects of signal to noise ratio of instrument and number of wavelengths on the accuracy of spectral analysis. Spectroscopy and Spectral Analysis, 2012, 32(8)： 2286-2289.
［8］Feng Z Q, Li F T. Design of a small sized fiber grating spectrograph measuring ultraviolet spectrum. Journal of Test and Measurement Technology, 2004, 18(4)： 292-295.
［9］Zhou Q T, Wei J, Xu Z P. Influence of noise feature on multiple sampling, accumulation and averaging technology. Infrared and Laser Engineering, 2010, 39(5)： 959-962.
［10］Wakabayashi H, Yamaguchi K, Yamagata Y. Noise reduction effect and analysis through serial multiple sampling in a CMOS image sensor with floating diffusion boost-driving. Japanese Journal of Applied Physics, 2017, 56： 04CF07.
［11］Li G, Zhou M, Lin L. Double-sampling to improve signal-to-noise ratio (snr) of dynamic spectrum (ds) in full spectral range. Optical & Quantum Electronics, 2014, 46(5)： 691-698.
［12］Suh S, Itoh S, Aoyama S, et al. Column-parallel correlated multiple sampling circuits for CMOS image sensors and their noise reduction effects. Sensors, 2010, 10(10)： 9139-9154.
［13］Nagura R. SN improvement ratio by time delay and integration for high-resolution Earth observation system. Electronics and Communications in Japan, Part 1 (Communications), 1995, 78(3)： 74-84.
［14］Welsh B M. Speckle imaging signal-to-noise performance as a function of frame integration time. Journal of the Optical Society of America A, 1995, 12(6)： 1364-1374.
［15］Zhan X X, Lin J, Zhou Z H. The technique of acquiring weak signal in NIR spectrometer. Chinese Journal of Scientific Instrument, 2002, 23(s1)： 29-31.
［16］Gobrecht A, Bendoula R, Roger J M, et al. Combining linear polarization spectroscopy and the representative layer theory to measure the beer-lambert law absorbance of highly scattering materials. Analytica Chimica Acta, 2015, 853(1)： 486-494.
［17］Gobrecht A, Bendoula R, Roger J M, et al. A new optical method coupling light polarization and Vis-NIR spectroscopy to improve the measurement of soil carbon content. Soil and Tillage Research, 2016, 155： 461-470.
［18］Han Y, Xu W R, Jin L. Study on the polarized reflectance hyperspectral characteristics and models of typical saline soil. Journal of Infrared and Millimeter Waves, 2014, 34(5)： 606-612.
［19］Zhang G J, Li L N, Li Q B, et al. Application of denoising and background elimination based on wavelet transform to blood glucose noninvasive measurement of near infrared spectroscopy. Journal of Infrared and Millimeter Waves, 2009, 28(2): 107-110.
［20］Fan Y C, Lu Q P, Ding H Q, et al. Study on the application of empirical mode decomposition to noninvasive hemoglobin measurement by NIRS. Spectroscopy and Spectral Analysis, 2013, 33(2)： 349.
［21］Zhuang X G, Wang L L, Wu X Y, et al. Origin identification of Shandong green tea by moving window back propagation artificial neural network based on near infrared spectroscopy. Journal of Infrared & Millimeter Waves, 2016, 35(2)： 200-205.
探究光谱仪信噪比对近红外光谱分析精度的影响
庄新港1,2,3, 史学舜1,2,3, 刘红博1,2, 刘长明1,2, 张鹏举1,2, 王恒飞1,2
（1. 中国电子科技集团公司第四十一研究所， 山东 青岛 266555； 2. 国防科技工业光电子一级计量站， 山东 青岛 266555； 3. 电子测试技术重点实验室， 山东 青岛 266555）
摘要： 信噪比作为光谱仪的重要指标之一， 体现了光谱仪对微弱信号的检测能力。 为了考察信噪比对光谱分析精度的影响， 首先介绍了影响光谱信噪比的主要因素， 然后以绿茶产地鉴别为例， 通过实验分析了光谱信噪比对绿茶产地溯源分析模型精度的影响， 同时拟合出光谱信噪比与光谱分析模型预测精度之间的关系， 讨论了提高光谱信噪比的方法。 结果表明， 随着光谱信噪比的降低， 训练集和预测集建模模型的准确度均呈现先缓慢下降， 然后直线下降， 最后趋于平缓的趋势。 通过计算可知， 为实现模型对预测集样本预测准确度达到90%和85%， 要求光谱信噪比分别高于23.42 dB和21.16 dB。 实验结果对新型在线分析用光谱类仪器的研制提供一定参数支持， 尤其是对仪器信噪比这一指标。
关键词： 近红外光谱； 信噪比； 偏最小二乘； 光谱分析； 绿茶
引用格式：ZHUANG Xin-gang, SHI Xue-shun, LIU Hong-bo, et al. Influence of signal-to-noise ratio on accuracy of spectral analysis by near infrared spectroscopy. Journal of Measurement Science and Instrumentation, 2020, 11（3）： 211-216. ［doi： 10.3969／j.issn.1674-8042.2020.03.002］
[full text view]
