基于近红外光谱技术的红小豆品质多指标无损检测及品种鉴别方法研究

Non-destructive Determination of Multiple Quality Indicators and Varieties of Red Bean Using Near-infrared Spectroscopy

本研究采用近红外光谱技术,构建8种不同品种红小豆的蛋白质、水分、支链淀粉、直链淀粉、百粒重、粒长、粒宽及颜色指标的同时快速无损分析方法。首先,对红小豆的近红外光谱进行标准正态变量变化法(SNV)、多元散射校正法(MSC)及一阶导数预处理,随后结合偏最小二乘分析法(PLS),建立红小豆样品不同品质指标的定量分析模型,结果显示:各PLS模型能较好地对红小豆的品质指标起到预测效果,达到快速无损检测红小豆品质的目的。此外,通过偏最小二乘判别分析法(PLS-DA)对红小豆样本的品种进行定性判别,校正集和预测集的判别正确率分别为100.00%和88.89%。近红外光谱对不同品种红小豆能做出较好的区分,对优质红小豆的快速筛分具有指导意义,且定性判别模型对同品种不同产地的2种红小豆同样得到了很好的区分,由此可见近红外技术在产地判别方面的可行性。

In this study,near-infrared spectroscopy was used to develop a rapid and nondestructive analytical method for protein,water,amylopectin,amylose,100-grain weight,grain length,grain width,and color indicators of eight different varieties of adzuki bean.First,the near-infrared spectra of the red bean samples were subjected to standard normal variable(SNV)modification,multivariate scatter correction(MSC),and first-order derivative preprocessing.Then,a quantitative analysis model for various quality indicators of red bean samples was constructed by combining partial least squares(PLS).The results indicated that each PLS model can effectively predict the quality indicators of red beans,achieving the goal of rapid non-destructive testing of red bean quality.In addition,qualitative analysis of eight kinds of red bean samples was carried out by partial least square discriminant analysis(PLS-DA),and the discrimination accuracy of the calibration and prediction sets were 100.00%and 88.89%,respectively.The model could discriminate different varieties of red beans well,crucial for rapid screening of high-quality red beans.In the qualitative discrimination model,two kinds of adzuki beans of the same variety from different countries of origin were also well discriminated,indicating the feasibility of near-infrared technology in origin discrimination.