不同产地黄硬皮马勃的红外光谱鉴别方法研究

Discrimination of Scleroderma flavidum from Different Geographical Origins by Infrared Spectroscopy

黄硬皮马勃（Scleroderma flavidum Ell.et EV.）主要生长在我国南方地区,有食用和药用价值。利用红外技术对其进行产地鉴别,为黄硬皮马勃的资源鉴别提供基础分析方法和理论依据。利用傅里叶变换红外光谱（Fourier transform infrared spectroscopy,FTIR）仪采集4个不同产地共40个黄硬皮马勃样本的红外光谱,每个样本平行扫描3次,取平均值。随机选择12个样本作为验证集,其余作为校正集,采用1 800～500 cm^-1波段的光谱数据,对比多种预处理方法,选择最佳预处理方法建立模型并进行判别分析（discriminant analysis,DA）。结果表明,预处理方法 ND（7∶3）＋SD＋MSC（ND为诺里斯导数平滑,SD为二阶导数,MSC为多元散射校正）构建的判别模型性能最佳,验证集样本的分类准确率和校正集样本的回判准确率均达到100%,模型鉴别效果良好,可靠性高。傅里叶变换红外光谱法结合判别分析,能有效鉴别不同产地黄硬皮马勃。

Scleroderma flavidum Ell. et EV. which has both medicinal and edible value,is mainly distributed in south China. Discrimination of S. flavidum from different geographical origins by infrared spectroscopy could provide an basic analysis method and theoretical basis for distinguishing its resource.Fourier transform infrared（ FTIR） spectroscopy method was used to study 40 S. flavidum samples from four different producing areas. Every sample was scanned for three times to get the average spectrum.Twelve samples were chosen randomly to form the validation set and others were used to build calibration set. According to the contrastive results of different pretreatment methods for the spectra with the range of1 800—500 cm^- 1,the best pretreatment method was used to build the discriminant analysis model. The results indicated that the model with pretreatment method ND（ 7 ∶ 3） ＋ SD ＋ MSC was the best. The prediction accuracy of the validation sets and the correct rate of returned classification of calibration set all reached 100%. The model had good effect and high reliability. FTIR combined with discriminant analysis could provide a fast and high-efficiency way to distinguish the S. flavidum with different geographical origins,effectively.