Study On Rapid Detection And Foliar Distribution Of Chemical Active Ingredients In Loquat (Eriobotrya Japonica) Leaves

Loquat(Eriobotrya japonica) leaf is an important health food resource in China.It is rich in chemical active ingredients with antioxidant, anti-cancer, anti-allergy effects. The content of chemical active components is an important index to judge the health function of loquat leaf. Traditional detection methods have high detection precision. But they cannot achieve rapid detection the chemical active ingredient of loquat leaf. At the same time, they have ignored the change of the chemical active components in loquat leaves with the change of growth stage. This leads to a low utilization of loquat leaves. In this paper, flavonoids, polysaccharides and triterpene acids contents in different growth stages(young leaf, mature leaf, old leaf) of loquat leaves will be detected using near-infrared spectroscopy and hyperspectral imaging technology. The main content of the research is as follows:(1) Acquisition of loquat leaves. To make the experimental data more representative, we pick leaves in February, May, August and November respectively,each month 10 young leaves, 10 mature leaves and 10 old leaves will be picked, then,a total of 120 leaves will be comprised. Select 80 as a correction set, 40 as the prediction set randomly.(2) Acquisition of near-infrared spectroscopy and hyperspectral images. Collect the near-infrared and hyperspectral images of each loquat leaf respectively, and spectral preprocessing. Then determination of flavonoids content in each sample by rutin standard method; determination of polysaccharides content of the samples by anthrone sulfuric acid method; determination of the triterpene acids content of the samples by ursolic acid standard method immediately.(3) Rapid determination by near-infrared spectroscopy. With the near-infrared spectroscopy collected by(2), the spectra will be selected to predict the flavonoids,polysaccharides, triterpene acids contents in loquat leaves. Respectively, using seven methods, i.e., partial least squares(PLS), interval partial least squares(i PLS),backward interval partial least squares(Bi-PLS), United interval partial least squares(Si-PLS), genetic partial least squares( GA-PLS), least square- support vectormachine(LS-SVM) and back propagation artificial neural network(BP-ANN), to build predictive models of three components, the best prediction model will be used to predict the loquat content of the components. The optimal model(RP=0.8776) for predication of total flavonoid was built using Si-PLS, while the optimal model(RP=0.8867) of polysaccharide was built using LS-SVM and the optimal model(RP=0.8508) of triterpene acid was built using LS-SVM. The spatial distribution of the principal components in the near infrared spectra showed that there were no significant differences in loquat leaves in different seasons, and significant differences in loquat leaves in different growth stages.(4) Rapid determination by hyperspectral imaging technology. With the hyperspectra collected by(2), the average spectra of region of interest(ROI) will be selected to predict the flavonoids, polysaccharides, triterpene acids contents in loquat leaves. Using seven methods above to build predictive models of three components,distribution maps of three components in three growth stages leaves(young leaf,mature leaf, old leaf) will be calculated based on each optimal models. The map shows a relatively low level of flavonoids and triterpene acids at the margins and some regions of blue mesophyll areas. Higher levels can be noticed in the regions along both sides of the veins and in some areas exhibiting yellow and jacinth tissue.The distribution of polysaccharide content is relatively uniform. Each components show obvious trends in different growth stages.The study has shown that using near infrared spectroscopy for chemical active ingredient quantitative prediction is instructive, distribution maps of three components in three growth stages leaves(young leaf, mature leaf, old leaf) have been calculated based on hyperspectral imaging technology. Each components show obvious trends in different growth stages. The study provides an effective reference for reasonable utilization of loquat leaves.