Estimation Model Of Rice Leaf Nitrogen Contet And Yield Based On Hyperspectral

In crop production,water and fertilizer are one of the most important limiting factor affecting the growth of crop.Besides,they are main components of crop organics(such as chlorophyll,nucleic acid and protein).Nitrogen plays an important role in the process of plant growth and is an important participant in photosynthesis.It is also an indispensable element in the process of plant growth and development.For a long time,the chemical analysis method is an important means to detect nutrient elements content in rice.But because of the complicated operation,inefficient conclusion and its destructive damage to crop,it cannot become a means of guiding fertilization in field.However,there are also disadvantages in undestructive fertilizer window method,leaf color card method and so on,such as hard to fertilize quantitatively and poor accuracy.In recent years,with the rapid development of hyperspectral technology,the use of hyperspectral technology for the diagnosis of nutrient elements in rice has become a new method of high speed,efficiency and undestruction detection.Obtaining and analyzing hyperspectral data of different rice varieties under different nitrogen treatment and different stages through field experiment in 2015,hyperspectral data that sensitive indicate rice leaf nitrogen content and yield are found,the inversion model based on hyperspectral parameters is constructed.It is tested by the data acquired in 2016.Specific research contents and achievements are as follows:(1)Through the field experiment in 2015 and 2016,Shendao 47 and Shendao 539 were cultivated under four levels of nitrogen treatment as experiment samples.(2)The hyperspectral characteristics of rice canopy and biophysical,biochemical parameters under different nitrogen levels and different growth stages are applied using airborne hyperspectral imaging system;The information of rice leaf spectral is obtained by spectrometer.The difference of canopy spectral response characteristics between Shendao 47 and Shendao 539 under different nitrogen levels and different growth stages are analyzed.At the same time,with the help of the analysis and test center of Shenyang Agricultural Uinversity,the physical and chemical parameters of rice were determined.Hyperspectral regression estimation models are constructed using these data.(3)Through the analysis on rice canopy spectral parameters,it is found that rice leaf nitrogen content has good correlation with Sr of mature stage(R2 = 0.8559*),DVI of heading stage(R2 = 0.7342*)and DVI of mature stage(R2 =0.8136*).Hyperspectral estimation model of rice leaf nitrogen content is constructed,the inspection accuracy is above 90%.(4)Through the analysis on rice canopy spectral parameters,it is found that rice theoretical yield has good correlation with Sr of heading stage(R2 = 0.932*),S= of mature stage(R2 = 0.775*),DVI of heading stage(R2 = 0.828*)and DVI of mature stage(R2 = 0.948*);rice measured yield has good correlation with Sr of heading stage(R2 = 0.874*),Sr of mature stage(R2 = 0.798*),DVI of heading stage(R2 = 0.815*)and DVI of mature stage(R2 = 0.917*).Hyperspectral estimation model of rice yield is constructed,the test accuracy is above 85%.