单粒糙米蛋白质含量的近红外分析数学模型

Mathematic Models for Analyzing Protein Content in Single Brown Rice with Near Infrared Spectroscopy

应用近红外透射光谱(NITS)技术,采用多种计量数学处理方法和不同的回归统计方法建立定标数学模型,优化得到了单粒糙米蛋白质含量测定的近红外定标模型。其定标相关系数(1VR)和检验决定系数(RSQ)分别为:0.924和0.940。内部交叉验证和外部验证结果表明,近红外定量分析有很高的准确度。

With the technique of near infrared transmittance spectroscopy （NITS）, the predicted models for protein content in single brown rice were established by several mathematics treatments, several different regression and then they were optimized. The regression squared（1 -VR）of the calibration and determination coefficients （RSQ）of the validation were 0. 924 and 0. 940 respectively. The method of NITS had high accuracy in predicted test of cross validation and independent validation and was available to replace the routine laboratory analysis. As a rapid and full measurement, the NITS method has potential application in rice breeding.