基于不同波段的何首乌多糖及总糖高光谱成像检测方法  

Detection method of hyperspectral imaging for polysaccharide and total sugar in Polygonum multiflorum based on different wave bands

作　　者：陈东[1] 顾玲玲[2] 

机构地区：[1]南京信息工程大学体育部,江苏南京210000 [2]江苏省人民医院肾内科,江苏南京210000

出　　处：《粮食与油脂》2018年第2期91-95,共5页Cereals & Oils

摘　　要：为实现何首乌内部品质的快速分级与评价,针对多糖和总糖含量两个何首乌品质指标,研究何首乌多糖和总糖含量与光谱间的关系,建立BP神经网络预测模型,探寻何首乌中多糖和总糖含量的无损检测方法。结果表明:基于全波段下的光谱信息的特征参量所建立的BP神经网络预测模型最佳,何首乌多糖含量预测的平均正确率为98.88%,相关系数平均值为0.996 2,均方根误差平均值为0.068 8;总糖含量预测的平均正确率达100%,相关系数平均值为0.993 5,均方根误差平均值为0.252 5。该研究为何首乌多糖和总糖含量的无损检测方法的实现提供理论依据,对实现何首乌内部品质的评定具有重要的现实意义。In order to achieve the rapid grading and evaluation of the internal quality of Polygonum multiflomm, the relationship between the polysaccharide and the total sugar content and the spectrum was studied, and the BP neural network prediction model was established for the polysaccharide and total sugar content of Polygonum multiflomm. The results showed that the prediction accuracy of the BP neural network was 98.88%, and the correlation coefficient was 0.996 2, and the mean square error was 0.068 8; The average correct rote of total sugar content was 100%, the correlation coefficient was 0.993 5, and the mean square error was 0.252 5. A theoretical basis for the realization of the non-destructive testing method ofpolysaccharide and total sugar content was provided, which was of great practical significance to the evaluation of the internal quality of Polygonum multiflorum.

关 键 词：高光谱图像 波段 BP神经网络 何首乌 多糖 总糖 

分 类 号：TS201.7[轻工技术与工程—食品科学]