莲子是我国重要的药食同源食物, 与莲子营养价值相当、 便于食用的莲子粉备受消费者青睐。 为保证莲子粉的品质, 利用近红外光谱(NIRs)技术对掺杂小麦粉、 玉米粉和地瓜粉的莲子粉进行鉴定, 在样品类别已知下利用支持向量机(SVM)、 最小二乘支持向量机(LS-SVM)、 偏最小二乘法-判别分析(PLS-DA)模型进行判别, 在样品类别未知下基于聚类算法进行判别。 同时, 对莲子粉中水分含量利用偏最小二乘(PLS)回归进行定量分析。 结果表明, LS-SVM模型对纯莲子粉样品与掺入小麦粉、 玉米粉和地瓜粉的莲子粉样品的判别率达到100%; 基于聚类算法能够有效识别掺入5%地瓜粉、 小麦粉和玉米粉的莲子粉样品; PLS模型对莲子粉中水分含量预测综合性能良好, 其中经过标准化预处理得到模型效果最佳, 其R2c, RMSEC, R2p和RMSEP分别达到0973 2, 0111 5, 0969 5和0118 9。 近红外光谱技术能为隐蔽的莲子粉掺杂的鉴别以及莲子粉中水分含量监控提供一种快速、 准确、 无损检测的分析方法, 为保证高档次莲子品质提供一种有益的思路。

Lotus seed is an important medicine and edible food, but to dry lotus seeds cook requires a long time, so lotus seed powder is more popular by consumers to adapt to the modern fast-paced way of life. In this paper, lotus seed powder adulterated with sweet potato powder, corn flour and wheat flour were identified by near infrared spectroscopy (NIRs) technique. Support vector machine (SVM), least squares support vector machine (LS-SVM) and partial least squares discriminate analysis (PLS-DA) were used to identify the model when thecategory was known, and the clustering algorithm was usedotherwise. In addition, the moisture content of lotus seeds powder was quantitatively analyzed by partial least squares (PLS) regression. The results showed that the discrimination accuracy of LS-SVM modelis 100%, and the clustering algorithm could effectively identify the 5% adulteration ofsweet potato powder, corn flour and wheat flour. Moreover, performance of PLS model to predict the moisture content in the lotus seed powder is good, and the accuracy of model by Normalize was satisfactory with the coefficients of determination of calibration (R2c=0973 2), the coefficients of determination of prediction (R2p=0969 5), root mean square errors of calibration (RMSEC=0111 5), and good root mean square errors of prediction (RMSEP=0118 9). The results showed that the near infrared spectroscopy is a fast, accurate and nondestructive analysis method to rapidly identify the lotus seed powder, accurately determinate the water content in lotus seed powder, and availably provide a useful idea for quality testing of daily food.