基于OSC-PLS可见光谱法检测白砂糖色值、混浊度及不溶于水杂质  

Rapid Determination of Color, Turbidity and Insoluble Matter Content in White Granulated Sugar by Visible Spectroscopy

作　　者：辛丽平 柴欣生[2] 周玉国[1] 田艳兵[1] 

机构地区：[1]青岛理工大学,山东青岛266033 [2]华南理工大学,广东广州510640

出　　处：《甘蔗糖业》2018年第1期50-57,共8页Sugarcane and Canesugar

基　　金：国家自然科学基金青年基金项目(21606141);山东省自然科学基金项目(ZR2016FB04)

摘　　要：提出了一种基于可见光谱(VIS)结合偏最小二乘算法(PLS)快速检测白砂糖色值、混浊度及不溶于水杂质的方法。该方法检测时无需调整pH=7,无需滤膜过滤处理,具有操作简单、分析快速的特点。为了提高检测精度,消除光谱中的噪声干扰,采用了一阶导数(FD)和正交信号校正(OSC)对原始光谱进行预处理。研究发现,选取350~360 nm及460~800 nm的特征波段,基于OSC光谱建立的关于白砂糖色值、混浊度及不溶于水杂质的多变量偏最小二乘回归模型(PLS2)具有很好的预测效果,对校正集和预测集样本的预测结果均令人满意,相关性系数都在0.96以上,相对均方根偏差都在10%以内。This paper reports on a spectroscopic method for rapidly determining the color value, turbidity and insoluble matter content of white granulated sugar. It was based on the spectral characterizations of these parameters in the visible wavelength range. Due to the complicity of the spectrum, a partial least squares regression （PLS） mode was used in the method calibration. The results showed that with the orthogonal signal correction （OSC） treatment on the primary spectrum, the PLS model could provide the best prediction for the color value, turbidity and insoluble matter content of the sugar samples, in which the correlation coefficients were better than 0.96 and all the relative root mean square deviations were less than 10%.

关 键 词：白砂糖 色值 混浊度 不溶于水杂质 OSC PLS 

分 类 号：TS247[轻工技术与工程—制糖工程]