Study On Modeling The Yield Estimation Of Camellia Oleifera Based On Spectral Characteristics

Camellia oleifera is a unique woody oil tree species in China.Studying the relationship between spectral characteristics and leaf nutrients is of great significance for accurate fertilization and yield estimation of Camellia oleifera plantations.Currently,the Camellia oleifera forests are mostly fertilized with soil testing formulas,which has a large workload and high cost,which is difficult to promote on a large scale.The output of Camellia oleifera is mostly harvested statistics,which is not conducive to operators to arrange management measures in advance.Remote sensing technology has been widely used in crop yield estimation,but there has not been any report on the estimation of oil tea production based on remote sensing.In this paper,the three Camellia varieties Pingtian Baisangyuan,Changlin No.40,and Dabieshan No.1 planted in Dechang Seed Seed Co.,Ltd.,Shucheng County,Anhui Province were taken as the research objects.Through the full-band ground spectrometer,plant canopy analyzer,The chlorophyll meter obtains the data of the individual plant spectrum,leaf area index,chlorophyll and nutrient content of the camellia for multiple months,analyzes the characteristics of the camellia spectral characteristics with time,and constructs a linear stepwise regression model of the spectral characteristic parameters and agronomic parameters of various varieties of camellia to reveal the spectral characteristics of the camellia In response to changes in agronomic parameters;at the same time,a model of the relationship between Camellia oleifera leaf area index and spectral characteristics was constructed;on this basis,a multi-spectral variable single-camera camellia yield estimation model was constructed.The project research can provide scientific basis for nutrient diagnosis of Camellia oleifera forest,precise fertilization,yield estimation,advanced management decision-making,etc.At the same time,it can also provide a research basis for parameter inversion of aerial remote sensing and ground remote sensing plant nutrient diagnosis and yield estimation.The specific research results are as follows:(1)The reflectance of various varieties of Camellia oleifera in the visible light and near infrared bands has certain regularity with the growth month.The green peak and red valley reflectance of Hirata Baisangyuan increased from July to September,and the near infrared reflectance increased in July and August,decreased in September,and increased in October.The green peak and red valley reflectance of Changlin No.40(born in 2012)began to decline from the rise in July and August to September,and the reflectance in the near-infrared band gradually decreased from July to October and reached the lowest value.Changlin No.40(born in 2015)and Dabieshan No.1 are very similar in the visible light slope and near infrared band.The reflectance of Changlin No.40(born in 2015)and Dabieshan No.1 in the visible and near-infrared bands has gradually increased from July,rapidly increased from August to September,and gradually decreased rapidly from October to near July and August Level.(2)Agronomic parameters and yield have a certain response to the spectrum.Pingtian Baisangyuan’s responsivity in the near-infrared band is higher than the chlorophyll content and the total nitrogen content of the leaves..The total nitrogen content of Changlin 40(born in 2015)is consistent with the monthly change of the spectral reflectance of Camellia oleifera.The responsivity of the phosphorus content of Dabie Mountain No.1 leaf and the reflectance in the near infrared band is good.Changlin 40(born in 2012)has higher reflectance than visible camellia in both visible and near-infrared bands.Changlin 40(born in 2015)The reflectance of the spectral curve with less production in July is much higher than that with more production,and the difference is higher than in other months.Dabie Mountain No.1 has the largest difference in reflectivity between July and October.The reflectivity in July with less output is higher than that with more output,and the opposite is true in October.(3)The yield of Changlin No.40(born in 2012)was extremely positively correlated with the leaf area index,and significantly negatively correlated with the total nitrogen content of the leaves.A model combining agronomic parameters and spectrum finally yields a composite model with the output of Changlin 40 as the independent variable and RVI,Rnir,NDVI,DVI,and DSI as the dependent variables:Vproduction=20.208+0.474X1+2.682X2-23.307X+5.474X4-27.7X5,where X1 is RVI,X2 is Rnir,X3 is NDVI,X4 is DVI,and X5 is DSI.The output of Changlin 40(born in 2015)is very significantly positively correlated with the leaf area index,and is significantly negatively correlated with the chlorophyll content.The model combining agronomic parameters and spectrum finally results in the output of Changlin 40(born in 2015)as an independent variable,RSI,RVI,and NDSI are dependent variable composite models:Vproduction=1.497+1.277X1+0.025X2-2.968X3,where X1 is RSI,X2 is RVI,and X3 is NDSI.The output of Dabieshan No.1 is significantly positively correlated with the leaf area index.The model combining the agronomic parameters and the spectrum finally results in a composite model with the output of Dabieshan No.1 as the independent variable and the RVI dependent variable:Vproduction=1.61+0.043X,where X is RVI.