涩柿可溶性单宁的可见/近红外漫反射光谱无损检测研究

旨在建立可见/近红外漫反射光谱与涩柿内部可溶性单宁之间的关系,以评价可见/近红外漫反射光谱在测量涩柿内部指标可溶性单宁的应用价值.在可见/近红外光谱区域(570～1848 nm),实验对比分析了不同数学建模算法、不同导数处理方法和不同散射及标准化处理的涩柿可溶性单宁定标模型.结果表明,应用改进偏最小二乘回归算法、一阶导处理和去散射处理所建涩柿可溶性单宁定标模型的预测性能较优,其定标交互验证相关系数(Rcv)和预测相关系数(Rp2)分别为0.722 7和0.678 5,定标交互验证均方根误差(RMSECV)和预测均方根误差(RMSEP)分别为0.148 4和0.176 3.研究表明,可见/近红外漫反射光谱对涩柿可溶性单宁的快速无损检测具有一定的可行性,但模型精度有待提高.

Research on nondestructive measurement of soluble tannin content of astringent persimmon using visible and near infrared diffuse reflection spectroscopy

The objectives of the present study were to establish the relationships of the visible and near infrared diffuse reflection (Vis/NIR) spectroscopy and the soluble tannin content of internal quality index of astringent persimmon, and to evaluate the value of Vis/NIR spectroscopy in measuring the soluble tannin content of internal quality index of astringent persimmon. In the spectral region between 570 and 1 848 nm, calibration results for the soluble tannin content of astringent persimmon were compared with different regression techniques, different derivative treatments and different scatter and standard treatments. The results showed that the modified partial least squares(MPLS) model, with respect to the first derivative D1 log(1/R) and detrend only, provided better prediction performance for the soluble tannin content of astringent persimmon fruit, with the correlation coefficient of cross validation of calibration (R(cv)) and correlation coefficient of prediction (R(p2)), the root mean square error of cross validation of calibration (RMSECV) and the root mean square error of prediction (RMSEP) of 0.722 7, 0.678 5, 0.148, and 0.176 3 respectively. The preliminary research on the built models indicated that nondestructive measurement of the soluble tannin content of astringent persimmon using Vis/NIR spectroscopy technique was feasible, but the precision of the models could be improved further.