论文部分内容阅读
目前南疆红枣在收购、清洗、加工和销售等环节上对水分的检测存在速度慢、精确度低、需破坏红枣等缺点,因此研究一种快速、简单、无损的红枣水分检测技术很有必要,近红外光谱技术对红枣水分检测简单、快速、无损。将新鲜的南疆红枣分成2组,一组用于建立红枣含水量校正模型,另一组作为验证模型,红枣的干燥方法采用标准烘干法。光谱数据预处理方法选择Savitzky-Golay导数和多元散射校正(MSC),校正模型的建立使用偏最小二乘回归分析法(PLS),所建模型SEC(校正集标准差)值达到最低为1.0667,RC(校正集预测值与真实值相关系数)与RPDC(性质值方差/SECV)达到最高值分别为0.98683和5.6258,主因子数为6,偏差(d)为-1.94~2.22,平均偏差(bias)为0.41%,极差(e)为2.22。结果表明模型可以对南疆红枣进行水分检测。
At present, there are some shortcomings in the detection, such as acquisition, cleaning, processing and marketing of jujube in south Xinjiang, so it is necessary to research a fast, simple and nondestructive jujube moisture detection technology. , Near infrared spectroscopy on jujube moisture detection is simple, fast, non-destructive. The fresh southern Xinjiang dates are divided into two groups, one for the establishment of the jujube moisture content correction model, the other as a validation model, jujube drying method using the standard drying method. Savitzky-Golay derivative and multivariate scatter correction (MSC) were selected for spectral data preprocessing. Partial least squares regression (PLS) was used to establish the calibration model. The SEC of the model was the lowest (1.0667) The correlation between RC (Correlation Set Prediction and Real Value Correlation) and RPDC (Nature Value Variance / SECV) reached the highest values of 0.98683 and 5.6258 respectively, the number of main factors was 6 and the deviation (d) was -1.94 ~ 2.22. The average deviation ) Was 0.41%, and the difference (e) was 2.22. The results show that the model can detect the water content of the southern red dates.