摘要
[Objective] This study was conducted to evaluate the feasibility of deter- mining crude protein in ramie using near-infrared （NIR） spectrometer. [Method] Par- tial least square regression （PLSR） was performed to establish a calibration model based on 50 samples for predicting the crude protein content in ramie, and the model was validated with data in the validation set consisting of 10 samples. [Result] The correlation coefficient of the model was 0.98. There was a good correla- tion between the predicted values by the near-infrared prediction model and the measured values by chemical analysis, and the relative error was 3.54% on aver- age between the predicted and the measured values. [Conclusion] The results showed that it is feasible to determine crude protein content in ramie using NIR spectroscopy-based prediction model.
[目的]探讨应用近红外光谱仪测定苎麻粗蛋白的可行性。[方法]以50个样品组成校正集,采用偏最小二乘法(PLS)建立近红外光谱信息与粗蛋白含量的校正模型,用该模型对10个样品的粗蛋白进行预测。[结果]该模型的相关系数为0.98。苎麻粗蛋白含量的化学测定值与近红外光谱模型预测值之间存在较好的相关性,预测值与化学值之间的平均相对误差为3.54%。[结论]用近红外光谱分析建立苎麻粗蛋白预测模型并测定苎麻粗蛋白含量是可行的。
