Study On Spectral Diagnostic Model Of Water And Nitrogen Of Summer Maize In North China

Irrigation and fertilization can promote the growth and development of summer maize to a certain extent,but unscientific irrigation and fertilization will cause resource waste and environmental pollution.The spectrum can accurately monitor the water and nitrogen status of crops,obtain the changes of physiological parameters and nutritional status of crops,judge whether they lack water and nitrogen,so as to timely supplement water and nitrogen content.Based on summer corn as the research object,this paper respectively in the national water saving irrigation engineering technology research center,Beijing daxing test research base and the sea of tranquility in tianjin area Tuan Bo Wa test base was carried out by field experiment,through the design different irrigation and fertilizer levels in summer maize growth period,carried out based on the technology of spectrum summer corn moisture and nitrogen diagnosis model research.The main research achievements are as follows:(1)SPAD value and plant nitrogen concentration can distinguish the nitrogen status of summer maize.The value of SPAD increased gradually from jointing stage to filling stage,and decreased after filling stage.The nitrogen concentration of plant decreased with the development of growth period.The spectral reflectance of visible band increases while that of near infrared band decreases.(2)The water content of summer maize was decreasing with the growth period.Blue and green wavelengths have low reflectance,while green and near infrared reflectance are high.(3)An empirical statistical model of aboveground nitrogen concentration was established based on sensitive band,hyperspectral index and new hyperspectral index.A multivariate prediction model was established based on the optimal new spectral index and above-ground nitrogen concentration.R~2 was 0.806,0.641,0.770,0.722,RMSE was 0.208,0.075,0.167,0.095,MAE was 0.178,0.063,0.122,0.081,respectively.(4)Based on the sensitive band,hyperspectral index and SPAD value,the prediction model of SPAD and aboveground nitrogen concentration was established.The optimization was based on the multivariate prediction model of hyperspectral index,SPAD value and above-ground nitrogen concentration.R~2 was0.500,0.524,0.566 and 0.531,RMSE was 0.149,0.170,0.147 and 0.154,respectively.MAE were 0.124,0.152,0.122 and 0.125,respectively.(5)By comparing the empirical model and PROSPECT model,the prediction model of sensitive band,hyperspectral index and aboveground nitrogen concentration was established with SPAD value as the intermediate variable.The results showed that the multivariate prediction model of hyperspectral index and aboveground nitrogen concentration based on PROSPECT model was better.R~2 increased from0.500 to 0.804 at jointing stage,from 0.524 to 0.846 at tasseling stage,from 0.566 to0.861 at filling stage,and from 0.531 to 0.554 at maturity stage.(6)The prediction model of multispectral band,multispectral index and plant water was established.The optimal bands at jointing stage,tasseling stage,filling stage and maturity stage were blue light,near infrared light,red light and green light,respectively.The optimal multispectral indices were OSAVI,NPCI,VARI and NPCI.Through comparative analysis,the correlation coefficients of multispectral index multivariate plant water prediction models were all above 0.5,RMSE and MAE were below 0.051 and 0.034,respectively.