在线检测白砂糖近红外光谱的预处理方法研究  

Study on near infrared spectrum pretreatment of on-line white granulated sugar determination

作　　者：王远辉[1] 黎庆涛[1] 姜毅[1] 黄康宁[1] 潘路路[1] 李震宇[1] 

机构地区：[1]广西大学轻工与食品工程学院,南宁530004

出　　处：《食品科技》2009年第8期283-286,共4页Food Science and Technology

摘　　要：为了找到在线检测白砂糖近红外光谱的最佳预处理方法,研究了标准化、均值中心化、Savitzky-Golay平滑、Savitzky-Golay求导、差分求导、标准正态变量变换、多元散射校正、静分析信号、正交信号校正、去趋势校正和基线校正等14种预处理方法的单方法和多方法联合处理,以交叉验证均方根和预测标准差作为评价指标。结果发现标准化、正交信号校正或静分析信号和1阶求导联合处理得到的效果最佳。To find the best near infrared spectrum pretreatment of on-line white granulated sugar determination, studying single and combine treatment of 14 pretreatments including autoscaling, mean centering, savitzky-golay smoothing, Savitzky-Golay derivative, difference derivative, SNV, MSC, NAS, OSC, DT and OFF etc. RMSECV and RMSEP are judgement index. Finally, the combination of autoscaling, NAS or OSC, 1 derivative is the best.

关 键 词：白砂糖 近红外光谱 预处理 

分 类 号：TS247[轻工技术与工程—制糖工程]