近红外光谱气体在线分析中基线校正方法的研究

光谱数据的预处理对于光谱定标及定量分析非常重要，尤其是校正测量条件、环境变化引起的光谱基线漂移是现场近红外光谱分析中的关键之一。利用BRUKER EQUINOX 55傅里叶近红外光谱仪对不同浓度的异丁烷气体的近红外吸收光谱进行测量，对 “微分-平滑”法和光谱仪自带的基线校正方法分别进行了理论分析和实验数据的对比。理论分析和实验结果表明，对于稳态、线性的背景光谱漂移光谱仪自带方法可以得到高的信噪比及分析精度，而“微分-平滑”法则具有更好的适应性，对于非稳态、或非线性漂移具有更好的抑制作用，是一种很好的在线数据预处理方法。

Investigation on Baseline Correction in Pollutant Monitoring On-Line by NIR Spectroscopy

Spectroscopic data pre-processing is important to spectra scaling and quantitative analysis, and correction of baseline drift caused by variation of measuring condition is critical in NIR spectroscopy analysis in-situ. In the present paper, the NIR absorption spectra of propane and isobutane with different concentration were detected by BRUKER EQUINOX55 spectrometer. The "Derivative-smooth" method and the baseline correction method which is accessory of the instrument were theoretically analyzed, and the data processed by the two kinds of method were compared. The results show that the method that is accessory of the instrument can correct steady-state and linear drift, and can gain high SNR, high accuracy; "Derivative-smooth" method has better applicability, can correct non-steady-state and nonlinear drift better, and is a good kinds of pre-processing method in analysis in situ.