Models Of Estimating Nutrient Elements In Leaves And Canopy Of Pyrus Sinkiangensis ’kuerlexiangli’ Using Hyperspectral Data

Pyrus sinkiangensis ‘Kuerlexiangli’ is a famous brand in Xinjiang characteristic forest and fruit industry.Its fruit sweet,thin crisp skin,mainly distributed around Kurla and Aksu area.In recent years,yellowing caused by lack of nutrient elements has been widespread in the main producing areas of Pyrus sinkiangensis ‘Kuerlexiangli’.When the Pyrus sinkiangensis ‘Kuerlexiangli’ is deficient in nutrient elements,the yield will not only decrease,but also the fruit will become smaller,the flavor will drop,the acidity will increase and the nutrient composition will decrease,which will seriously affect the fruit quality.Therefore,it is very important to estimate the nutrient element content of fragrant pear by spectrum method in real time,fast and high efficiency.In this study,Korla fragrant pear in Aksu Hongqipo Farm was used as the research object.ASD Field Spec 4 spectrometer was used to determine the leaf of fragrant pear and canopy spectral reflectance in different periods.At the same time,the leaves were collected in the range of canopy spectral measurement,and the content of nutrient elements was determined indoors.The variation rules of leaf and canopy spectra and nutrient elements in different periods of fragrant pear were analyzed.Spectral differential technique was used to screen spectral sensitive indicating wavelengths and a spectral estimation model based on stepwise regression analysis for nutrient elements was established.The main findings are as follows:The spectral characteristics and nutrient contents of leaves and canopy of fragrant pear indicated that: The spectral reflectance of fragrant pear leaves in the visible light band(400～700 nm),near infrared band(700～1 300 nm)and middle infrared band(1 300～2 500 nm)were lower than other periods in June.In this range,the spectral reflectivity in July was the highest,and the spectral reflectivity in August and September showed a decreasing trend with the passage of time.However,the canopy spectral curve of fragrant pear gradually decreased with the passage of time.It can be seen that the variation characteristics of canopy spectral curve of fragrant pear leaves from July to September are basically the same,and gradually decrease with the passage of time.The change of nutrient element content is different in different period,and July is the node of nutrient element content change.The correlation between nutrient elements and leaf and canopy spectra of fragrant pear was analyzed.The correlation between the original spectra of fragrant pear and nutrient elements was weak in June to September,and the correlation between the first order differential spectrum and nutrient elements was stronger.The absolute value of the maximum correlation coefficient between leaf spectrum and nutrient elements in different periods is between 0.512～0.921.The absolute value of the maximum correlation coefficient between the canopy spectrum and nutrient elements in different periods is between 0.325～0.820.The sensitive bands of canopy spectra were mainly concentrated in the near infrared(700～1 300 nm)and mid infrared(1 300～2 500 nm)range.The independence test of nutrient element estimation model in leaves and canopy of fragrant pear was carried out,The spectral band bivariate corresponding to the large correlation coefficient of the first order differential spectrum and the nutrient element is chosen as independent variable,and the nutrient elements in different periods are taken as dependent variables,and the modeling method of stepwise regression analysis is adopted.A regression model for estimating the content of nutrient elements in leaves and canopy of fragrant pear at different stages was established.The fitting degree of spectral nutrient element spectral estimation model of fragrant pear leaves in different periods was between 0.344～0.881,Best fitting of K,Fe,Mn in June Model,Best fitting of Ca in July Model,The best fitting of N,P,Mg,Cu,Zn element in September model was obtained.The fitting degree of the canopy spectral model was between 0.252 ～0.714.K,Ca,Fe,Mn,Cu element fitted best in June model,P element fitted best in July model.Optimum fitting of Mg,Zn in September Model.