论文部分内容阅读
目的:建立滇重楼生长年限的鉴别方法,为滇重楼的质量控制提供参考。方法:利用傅里叶变换红外光谱技术测试了同一产地7种不同生长年限的68株重楼主根木质部和表皮样品的红外光谱,利用Omnic8.0软件计算样品的二阶导数光谱和四阶导数光谱,分别以主根木质部和表皮在1 800~900 cm-1的红外光谱、二阶导数红外光谱和四阶导数红外光谱数据为样品特征,采用Mahalanbis距离逐步判别分析法,依据Fisher线性判别准则建立重楼生长年限的判别模型,对未知生长年限的样品进行鉴别比较。结果:四阶导数光谱比其他级别的光谱鉴别效果好,表皮光谱比木质部光谱的鉴别效果好;利用表皮的四阶导数光谱数据建立判别分析模型对重楼的生长年限进行鉴别时,训练样本回判正确率100%,测试样本预测正确率78.6%,总正确率95.6%。结论:傅里叶变换红外光谱结合逐步判别分析法可鉴别重楼的生长年限,为生产实践中预测重楼的生长年限提供了一种新思路。
OBJECTIVE: To establish the identification method of the growth years of Polygonum grandiflora in Yunnan and provide reference for its quality control. Methods: Fourier transform infrared spectroscopy was used to test the infrared spectra of the xylem and epidermis of 68 rootstocks of seven reeds of the same producing area. Omnic 8.0 software was used to calculate the second derivative spectra and the fourth derivative spectra , The infrared spectra, the second derivative infrared spectra and the fourth derivative infrared spectra of the xylem and epidermis of the main root were used as the sample features respectively. The Mahalanbis distance stepwise discriminant analysis method was used to establish the model according to the Fisher linear discriminant criterion Building growth discriminant model, the unknown growth period of the sample were identified and compared. Results: The fourth-order derivative spectrum is more effective than the other spectra, and the epidermis spectrum is better than the xylem spectrum. By using the fourth-order derivative spectral data of the epidermis to establish the discriminant analysis model to identify the growth period of the re-building, The accuracy rate of sentence was 100%, the prediction accuracy of test sample was 78.6%, and the total correct rate was 95.6%. Conclusion: Fourier transform infrared spectroscopy combined with stepwise discriminant analysis can identify the growth of re-building years, and provide a new idea for predicting the growth of re-building in production practice.