Wheat Canopy Spectral Features And Its Research Relationship With Agronomy Parameter Under Different Nitrogen Levels

Wheat is one of the world’s most widely distributed and have largest planting area food crops, who’s planting area and output only second to rice in our country. In addition, nitrogen has the most outstanding effect on wheat growth, development and production quality. Therefore, this paper through research wheat canopy spectral features and its research relationship with agronomy parameter under different nitrogen levels, provides theoretical basis and technical support for wheat growth remote sensing monitoring and production evaluation in Shaanxi guanzhong regionThis study used the SVC HR-1024 i portable spectrometer produced by The Spectra Vista Company in America to monitoring wheat canopy spectral reflectance, synchronous acquisition agronomy parameter. By studying the wheat canopy spectral characteristics under different nitrogen levels, the correlation between high canopy spectra and its derivative spectrum parameters with leaf nitrogen content, plant moisture content, chlorophyll and fresh biomass, adopting the method of stepwise regression, this paper established wheat high spectral estimation model of agronomy parameters, then tested it. The main research results are as follows.(1) In different growth period(N3 level), the change trend of wheat canopy spectral reflectance is consistent. Under different nitrogen levels(take green period for example), wheat canopy spectral reflectance performed differences. With the increase of nitrogen level, wheat canopy spectral reflectance gradually raised. In different nitrogen level and different growth period, the red edge of wheat has a "twin peaks" phenomenon. Under different nitrogen levels, the difference change of λr is small, and performance is not obvious. The difference change of Dr and SDr is big, performance is obvious. With the increase of nitrogen level, Dr and SDr kept rise first and then felled. In the reproductive period, with the development of growth period, λr, Dr and SDr all have trend to move to the red edge, reached the maximum in the filling stage. After milk stage moved to the blue edge direction.(2) Through analyzed the correlation between wheat leaf nitrogen content with canopy spectra, derivative spectra and spectral parameters, this paper got that: leaf nitrogen content was negatively related to the visible region, positively related to near infrared region(730-1400 nm). After 1400 nm, the correlation coefficient began to decline until the negative, and performed volatile. At the same time this paper got canopy spectral band, derivative spectral band and high spectral parameters which was correspond with the maximum correlation coefficient. Used these parameter to establish the high spectral estimation model of leaf nitrogen content and then texted it. Finally selected ND [1363, 622] as estimating model of leaf nitrogen content. Equation was: Y= 3.5126X1.4336, R2=0.761, RMSE=1.274, RE=23.64. Using the field experimental data in Shaanxi Rougu and Juliang to inspection the established model of wheat leaf nitrogen content, the results showed that estimating model of wheat leaf nitrogen content has good accuracy and universality.(3) In different growth period, plant water content were different from each other, but the difference were small. Under different nitrogen levels, change of plant water content was not obvious. Through analyzed the correlation between plant water content with canopy spectra, derivative spectra and spectral parameters, this paper got that: plant water content was negatively related to the visible region, positively related to near infrared region(760-1150 nm). After 1150 nm, the correlation coefficient began to decline until the negative, and performed volatile. At the same time this paper got canopy spectral band, derivative spectral band and high spectral parameters which was correspond with the maximum correlation coefficient. Used these parameter to establish the high spectral estimation model of plant water content and then texted it. Finally, selected Dr as estimating model of leaf nitrogen content. Equation was: Y=-10436X2+189.67X-0.0432,R2=0.766,RMSE=0.053,RE=5.34.Using the field experimental data in Shaanxi Rougu and Juliang text field to inspection the established model of plant water content, the results showed that estimating model of plant water content has good accuracy and universality.(4) In green stage and jointing stage, wheat chlorophyll content was higher, declined in heading stage, picked up in the filling stage. Then it continue to fall, until to ripening period. Through analyzed the correlation between chlorophyll content with canopy spectra, derivative spectra and spectral parameters, this paper got that: chlorophyll content was negatively related to the visible region, positively related to near infrared region(745-893 nm). After 893 nm, the correlation coefficient began to decline until the negative, and performed volatile. At the same time this paper got canopy spectral band, derivative spectral band and high spectral parameters which was correspond with the maximum correlation coefficient. Used these parameter to establish the high spectral estimation model of chlorophyll content and then texted it. Finally selected NDVI [773,590] as estimating model of leaf nitrogen content. Equation was: Y=-370.36X2+667.34X-251.3,R2=0.523,RMSE=9.24,RE=27.82.Using the field experimental data in Shaanxi Rougu and Juliang to inspection the established model of chlorophyll content, the results showed that estimating model of chlorophyll content has good accuracy and universality.(5)Under different nitrogen levels and in different growth period, the changes of fresh biomass was increased at first and then falled. Through analyzed the correlation between wheat fresh biomass with canopy spectra, derivative spectra and spectral parameters, this paper got that: leaf nitrogen content was negatively related to the visible region, positively related to near infrared region(730-1100 nm). After 1400 nm, the correlation coefficient began to decline until the negative, and performed volatile. At the same time this paper got canopy spectral band, derivative spectral band and high spectral parameters which was correspond with the maximum correlation coefficient. Used these parameter to establish the high spectral estimation model of fresh biomass and then texted it. Finally selected NDVI [801,670] as estimating model of leaf nitrogen content. Equation was: Y=-190.64X2-316.56X+134.88,R2=0.484,RMSE=0.745,RE=15.32.Using the field experimental data in Shaanxi Rougu and Juliang to inspection the established model of fresh biomass, the results showed that estimating model of fresh biomass has bad accuracy and universality.