【摘要】 水稻稻草硅质化是影响稻谷生产与稻草营养品质的主要因素之一 ,对水稻稻草硅化特点的化学分析费时、耗资、准确度较低。利用近红外漫反射光谱法 (NIRS)具有分析速度快、精度高、结果稳定的优点。对同年采自福建省不同种植地、不同季节、不同品种的稻草及稻叶、叶鞘、茎秆共 92份样品进行半纤维素、纤维素、木质素、可溶及不可溶性硅化物成分的近红外光谱分析。研究利用改进最小偏差 (改进最小二乘法回归技术 )回归法 (ModifiedPLS)进行定标建模 ,并进行交叉验证 (cross validation) ,探讨了用近红外漫反射光谱技术建立稻草纤维及硅化物组成快速预测模型的可行性 ,以及影响NIRS技术在建模应用中的主要因素。更多还原

【Abstract】 The silicification of rice straw is a factor that affects the grain production and straw nutritive quality. The procedure of chemical analysis on silicon in straw is, however, time and labor consuming, and slightly poor in accuracy. The study has attempted to apply near infrared reflectance spectroscopy (NIRS) technique as an advanced alternative to predict the fiber composition and silicification of rice straw. Ninety-two samples from different seasons and varieties were collected over the provincial region. Their chemical analyses were carried out on the aspects of hemicellulose, cellulose, lignin, extractable and non-extractable silicon, and the results were used as a data base for NIRS analyses. The factors that might affect the calibration, cross-validation and the prediction for the application of NIRS on rice straw were also discussed.更多还原