采用近红外光谱技术结合化学计量学算法,建立了成品葡萄酒(总糖和酒精度)的快速检测方法.通过NCL对近红外光谱图谱进行预处理,使用偏最小二乘法(PLS)建立了总糖和酒精度近红外定量模型.以决定系数(R2)、校正均方根误差(RMSEC)和预测均方根误差(RMSEP)为模型质量的评价指标.R2分别为0.9684、0.9146,RMSEC值分别为0.0810、0.2382,RMSEP值分别为0.0877、0.2352.结果表明,建立的模型预测效果较佳,具有较好的稳定性和较高的精密度,能满足成品葡萄酒生产中总糖和酒精度的快速检测需求.%A method of near infrared spectroscopy coupled with chemometric algorithm had been developed for rapid determination of total sugar and alcohol content of grape wine. The spectrum were pretreated by NCL, and NIR models of total sugar and alcohol content were constructed by partial least square (PLS). The correlation coefficient of calibration (R2), the root mean square error of calibration (RMSEC) and the root mean square error of prediction (RMSEP) were used as the model evaluating indexes. R2 values were 0.9684 and 0.9146 respectively, RMSEC values were 0.0810 and 0.2382 respectively, and RMSEP values were 0.0877 and 0.2352 respectively. The results demonstrated the satisfactory prediction effects of the established models. Besides, it had good stabili-ty and high precision and it could meet the requirements of rapid determination of total sugar and alcohol content in wine.
展开▼
