红外光谱结合判别分析对三七种植土壤类型的鉴别研究

Identification of Planting Agrotype of Panax notoginseng by Infrared Spectroscopy Combined with Stepwise Discriminant Analysis

利用傅里叶变换红外光谱（FTIR）技术结合判别分析,对三七（Panax notoginseng）的种植土壤类型进行鉴别研究。测试了6种种植土壤类型共102株三七植株主根木质部和须根样品的红外光谱;以1 800~900 cm-1的光谱信息为变量,每种土壤类型任选5份样品为测试样本,其他为训练样本;利用SPSS 18.0统计分析软件中的判别分析模块依据Fisher线性判别准则,采用逐步判别分析中的5种挑选样本信息变量算法,建立分类模型,对三七的种植土壤类型进行鉴别研究。结果表明,基于三七须根样品信息的种植土壤类型的判别结果明显优于基于主根木质部的判别结果,且5种挑选变量的算法建立的模型都能识别三七的种植土壤类型;综合考虑判别分析结果和模型的稳定性,＂Smallest F ratio＂法所建立的模型更适合用于三七种植土壤类型的鉴别。

A method based on Fourier transform infrared spectroscopy（FTIR） combined with stepwise discriminant analysis was used to distinguish the planting agrotypes of Panax notoginseng. The spectra in the range of 1 800~900 cm-1of xylems of taproot and fibril of 102 P. notoginseng plants from 6 planting agrotypes was analyzed using the discriminant model based upon the Fisher linear discriminant criterion. The discrimination effectiveness was compared for the part of plants and the choice among the 5 discriminant indexes in the discrimination function. The results indicated that the models based on the spectra of fibril showed better recognition effect than that of xylems of taproot, and synthesizes discrimination effectiveness and the stability of model, the Smallest F ratio method of stepwise discriminant analysis should be better for recognizing of planting agrotypes of P. notoginseng.