论文部分内容阅读
该研究对采摘于新疆阿克苏十团枣园的9月、10月、11月份3个月的骏枣进行了近红外光谱测定和总糖测定。结果表明,阿克苏骏枣总糖含量的测定采用PLS所建模型预测效果最佳,但最佳预处理方法不同;白熟期骏枣糖含量的最佳模型是经一阶导数处理后对骏枣糖度预测效果最优,校正相关(Rc)为0.985 8,RM-SEC为0.650,RMSEP为2.01,预测相关系数(RP)为0.903 4;脆熟期、完熟期骏枣糖含量的最佳模型都是经二阶微分处理后对骏枣糖度预测效果最优,校正相关系数(Rc)分别0.961 3、0.972 4,RMSEC分别为0.801、1.30,RMSEP分别为0.944 2、1.000 0,预测相关系数(RP)分别为2.90、1.56。该试验所建近红外PLS校正模型具有较好的稳定性,能满足红枣总糖含量的检测要求。
In this study, the Jujube jujube collected in September, October and November in Zaoyuan of the ten groups in Aksu, Xinjiang was measured by near infrared spectroscopy and total sugar. The results showed that the determination of the total sugar content of Aksu Jujube using PLS model to predict the best, but the best pretreatment methods are different; white jujube jujube sugar content of the best model is the first derivative of the jujube The optimal prediction of the brix was 0.985 8 for calibration, 0.650 for RM-SEC, 2.01 for RMSEP and 0.903 4 for the prediction of root mean square (Rc) were 0.961 3,0.972 4, RMSEC was 0.801,1.30 respectively, RMSEP was 0.944 2,1.000 0 respectively, and the predictive correlation coefficient (RP) ) Were 2.90,1.56 respectively. The near-infrared PLS calibration model built in this experiment has good stability and can meet the detection requirements of the total sugar content of jujube.