Study On Identification Of Polygonatum Kingianum Based On Multispectral Combination With Chemometrics

Objective Polygonatum kingianum is a plant of the same origin of medicine and food,and it has a wide distribution range.Different producing areas and growth years make the effective components of P.kingianum vary greatly.There is also a lack of unified standards for processing methods,and the plants of the same genus are highly similar,resulting in uneven quality.Methods 1.The polysaccharide contents and 13 agronomic traits from 2 to 5 years of P.kingianum were determined to explore whether the traits had influence on the accumulation of polysaccharides.2.Extracting the two feature information of the ATR-FTIR and UV-Vis spectra of P.kingianum and conduct mid-level data fusion to establish partial least square discriminant analysis model,so as to identify the samples of P.kingianum in different years.3.Random forest algorithm was used to compare the identification effects of single spectra and low-,mid-and high-level data fusion strategies on the origin of P.kingianum.4.Determination of polysaccharides in different processed products and content of 5-HMF produced during processing.Near-infrared spectroscopy and UV-Vis spectroscopy were used in combination with principal component analysis to explore the relationship between different processed products.5.ATR-FTIR combined with partial least squares discriminant analysis,orthogonal partial least squares discriminant analysis and support vector machine algorithm were used to identify the genus Polygonatum.Results 1.Polysaccharide content of P.kingianum extract varies greatly with different growth years,and the best harvest time is the 4th year after planting;There was significant correlation between the fibrous root number and the accumulation of polysaccharide.Feature extraction is an effective method to identify different growth years,and the identification accuracy is 100%.2.The identification effect of data fusion is better than that of single spectra;The identification effect of mid-level data fusion is better than that of high fusion and low fusion.The identification effect of latent variables is better than principal components.3.NIR and UV-Vis fingerprint spectra showed differences in the spectra of different processed products,and 4400-4200 cm^-1 and 220-400 nm were the fingerprint regions.4.The combination of latent variables and support vector machine algorithm can realize the identification of polygonati rhizoma with subtle differences.Conclusion 1.The best harvest time of P.kingianum is the 4th year after planting.Inhibiting the development of fibrous root can promote the accumulation of polysaccharide.2.Data fusion strategy is an effective method to improve the identification effect,among which feature variable extraction is the best fusion method.3.In the process of processing,the steaming time had a greater impact on the Polygonati Rhizoma,and the NIR and UV-Vis spectra were also changed to some extent.4.Spectrum combined with chemometrics can be used to identify the genus Polygonatum accurately.