可见光-近红外光谱预测粳稻食味的适应性研究

Predicting Eating Quality of Short Grain Rice by Using Visible and Near Infrared Spectroscopy

利用黑龙江省、吉林省、辽宁省和江苏省的88份常规粳稻品种,建立了适合中国粳稻食味评价的可见光-近红外光谱预测模型。感官食味试验采用小样品感官试验综合评分法,其相关性分析和逐步回归结果表明,感官综合评分与气味、色泽、外观完整性、味道、口感均呈极显著正相关。通过粳稻感官食味评分值与可见光-近红外光谱值的回归分析,得到7个自变量（4个波长的吸光度、碎米率、裂纹米率和香气系数）的预测回归方程,其相关系数R均达到0.699~0.799的极显著相关,经验证该预测回归方程能够较好的预测粳稻的米饭食味品质。

Using samples of 88 short grain rice varieties from Heilongjiang,Jilin,Liaoning,and Jiangsu provinces,a visible and near infrared spectroscopy prediction model for short grain rice taste evaluation was built up.The small sample sensory integrated scoring method was used in taste experiment,and correlation analysis and stepwise regression were carried out.Results show the sensory taste score is highly positively correlated with smell,luster,grain suit,taste,and texture.Through the regression analysis of short grain rice taste evaluation value and visible and near infrared spectroscopy value,a predicting regression equation based on 7 independent variables（absorbance of four different wavelengths,broken rice rate,crackle rice rate,and aroma coefficient） was obtained.Results show the correlation coefficients（R） reach 0.699~0.799,indicating highly significant correlation.It is confirmed that the prediction equation can predict the eating quality of short grain rice quite well.