应用近红外透射光谱法测定稻米淀粉回复值

摘要

Two hundred and thirty-five samples of rice were scanned by a near infrared transmittance spectrometer (NITS) , and a calibration equation was established by using modified partial least square method. The standard error of calibration (SEC), the standard error of cross-validation (SECV), and the reqression squared quotient (RSQ) of the calibration were 14. 28, 14. 29 and 0. 87, respectively. The results of internal cross-validation showed that the corrected standard error of performance (SEP), the average bias of the mean between reference values (BIAS), and the one minus variance ratio (1 -VR) of cross validation were 5. 58, 0. 40 and 0. 95, respectively. The results of internal cross-validation and external validation show that the NITS has high accuracy in determination of the viscosity characteristic value of rice starch.%以235份精米为样品,利用近红外透射光谱分析仪扫描样品,采用改进的偏最小二乘法建立定标模型,得到了近红外分析模型的定标标准偏差(SEC)、交叉检验标准差(SECV)和定标决定系数(RSQ),分别为14.28、14.92和0.87.内部交叉验证结果显示,经校正的工作标准差(SEP)、检验偏差(BIAS)和交叉检验相关系数(1-VR)分别为5.58、0.40和0.95,相关性达极显著水平.内部交叉验证和外部验证结果表明,近红外定量分析精米淀粉回复值有很高的准确度.