南疆骏枣叶片含水量的近红外光谱检测研究

Determination of Water Content in Leaves of Jun Jujube in Southern Xinjiang by Near Infrared Spectroscopy

[目的]建立南疆骏枣叶片含水量的快速、无损的检测模型。[方法]选取73片完好、无损的骏枣叶片,运用NIR检测骏枣叶片含水量的重要指标。通过3种不同的光谱预处理方法进行预处理,建立骏枣叶片含水量的PLS检测模型。[结果]在预测骏枣叶片含水量的PLS模型中,最好的组合是原始光谱+MSC+PLS,相关系数(R)由原始的0.673 1提高到0.874 6,预测精度(Precision)由0.950 7提高到0.957 8,预测残差平方和(PRESS)由0.028 4降低到0.017 7,预测标准偏差(RMSEP)由0.037 7降低到0.029 7。[结论]应用NIR技术不仅对南疆骏枣叶片含水量的快速、无损检测具有可行性,同时还对其他农作物叶片水分、叶绿素、氮含量光谱预处理检测具有一定的借鉴意义。

[Objective] A rapid and non-destructive detection model of la moisture content in Jun jujube of Southern Xinjiang was estab-lished. [Method ] 73 pieces of intact and undamaged Jun jujube leaves were selected, and NIR content of Jun jujube leaves. 3 different spectral pretreatment methods were applied to establish PLS detection model for water content of Jun jujube lea . [ Result] The best combination in the PLS model predicting the water content of Jun jujube leaves was the original spectrum ＋MSC ＋PLS , the correlation coefficient （R ） was rised from 0.673 1 to 0.874 6, the 0.957 8,and the predicted residual square sum （PR ESS） was reduced from 0.028 4 to 0.017 7 ,and the prediction standard deviation （RM-SEP ） was reduced from 0.037 7 to 0.029 7. [ Conclusion ] The application of NIR technology was not only feasible for the rapid and non-destruc-tive testing of water content in the leaves of Southern Xinjiang jujube, but also provided certain reference significance for the spectral pretreat-ment and detection of water, chlorophyll and nitrogen content in other crops.