基于UVE-PCA的两级三维荧光光谱波长选择方法

针对三维荧光光谱仪的传统分析手段不适于现场快速测量的特点,在分析传统红外二维波长选择方法的基础上,提出了基于UVE-PCA的两级准三维荧光光谱波长选择方法。通过UVE初选和PCA终选,在保证回归精度的前提下最大限度地压缩波长变量,实现三维荧光光谱的在线检测。使用丹麦数据库中的三维荧光光谱数据进行验证,实验表明经该方法筛选后的激发-发射波长序列压缩为全波长序列的1.86%时,模型预测能力提高了19.7倍,预测速度提高了7.6倍。该方法对三维荧光光谱在现场快速定量测量的实现提供了依据。

The Selection Method of Two-stage Three-dimensional Fluorescence Spectrum Wavelength Based on UVE-PCA

Aiming at the traditional analysis of three-dimensional fluorescence spectrum instrument not suitable for the characteristics of the fast measure at scene,based on the analysis of the traditional infrared wavelength selection method of Two-dimensional,this article put forward a selection method of two-stage three-dimensional fluorescence wavelength based on UVE-PCA.By primary selection of UVE and final selection of PCA,to furthest compress wavelength variables on the premise of regression accuracy and realize on-line detection of three-dimensional fluorescence spectrum.The article validate the data of three-dimensional fluorescence spectrum in Denmark database,when inspire-emission wavelength sequence by this method screening is compressed to 1.86% of whole wavelength sequence,the model prediction capability is improved by 19.7 times,forecasting,speed is improved by 7.6 times.The experimental results show that this method provides practical value for rapid quantitative measure three-dimensional fluorescence spectrum on site.