麻疯树种子含油量近红外光谱定标模型的建立

Calibration model building for investigating seed oil content of Jatropha curcas using near infrared spectroscopy

【目的】本研究旨在建立麻疯树种子含油量的近红外光谱定标模型,为近红外光谱技术应用于麻疯树种子含油量的测定提供依据。【方法】利用瑞典波通DA7200型近红外光谱成分分析仪采集了125份麻疯树种子样品的光谱数据,通过样品化学值测定、光谱数据预处理以及回归统计方法建立了麻疯树含油量的近红外光谱定标模型,并利用10份未知含油量的种子样品对模型的准确性进行了检验。【结果】本研究建模所选125份样品的含油量范围为25.23%~39.73%,平均值为33.91%±2.64%,中位数为34.31%,基本覆盖了当前麻疯树的主要品种,在种子油含量上具有一定的代表性。二阶导数联合标准正态变量转换法为建立麻疯树种子含油量近红外光谱定标模型的最佳预处理方法,偏最小二乘法为最佳的回归方法。模型验证结果显示,样品的化学测定值与近红外光谱定标模型预测值极显著相关,且相关系数达0.955 6,预测标准偏差为0.653 6。【结论】检验结果说明该模型具有较高的可靠性,可应用于后期麻疯树种子油分含量的大批量快速测定。

[ Objective] In present study, a calibration model was established to investigate the Jatropha curcas seed oil content using near infrared spectroscopy. [ Method ] The spectroscopic data of 125 Jatropha curcas seed samples was collected by a near infrared reflectance spectrometer DA7200 （designed by Perten）. With the chemical test method, spectral data pre-treatments and different regression methods, a calibration model using near infrared spectroscopy for seed oil content determination was established. And then the accuracy of the model was tested by 10 validation samples. [ Result] The seed oil content ranged from 25.23% to 39.73% among 125 investigated samples in this study, mean value was 33.91% _＋ 2.64% , and the median was 34.31%. It indicated that these samples could cover the main cuhivars of Jatropha curcas in the wild. Furthermore, second-derivative combined with standard normal variable transformation method was the best pre-treatment way in calibration model building for investigating seed oil content of Jatropha curcas using near infrared spectroscopy, and the partial least squares was the best regression method. Moreover, the significant correlation was obtained between the predicted value of near infrared spectroscopy calibration model and the chemical determination value of sample, and the coefficient was O. 955 6, while the standard deviation of prediction was O. 653 6. [ ConclusionI Our prediction model has high reliability, which can quickly investigate the seed oil content of Jatropha curcas in large quantities.