摘要
本文介绍了滤光片型近红外光谱茶叶成分分析系统的应用软件研制。软件具有光谱采集、数字信号处理、光谱预处理、模型定标、样品预测和系统自检等功能,并集成了对连续光谱的遗传算法、变量标准化及去势处理算法,以确定检测对象主成分的特征波长,为仪器的光源和分光系统的设计提供依据,以增强仪器的可扩展性及检测精度。采用多元线性回归建立了茶多酚、氨基酸、水分、咖啡碱等四种成分的定量分析模型,实现了基于近红外光谱技术的茶叶主要成分的快速无损检测。
A software was developed for tea content analysis based on filter near-infrared spectrometry. The software involves spectrum acquisition, digital signal processing, spectrum pretreatment, modeling, prediction, system self-checking and etc., and integrates genetic algorithm, standardized normal variate and de-trending to determine the characteristic wave lengths. The basis for designing light source and optical splitting system is given, which enhances the expandability and test precision of the system. The quantitative analysis models of tea polyphenol, amino acid, moisture content and caffeine are established using multiple linear regression method. The system achieves fast and non-destructive inspection for the main contents of tea.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2009年第11期2411-2416,共6页
Chinese Journal of Scientific Instrument
基金
浙江省科技计划项目(2006C2144)资助
