Research On Diagnosis Of Nitrogen Status In Maize Based On Leaf Spectral Analysis

The application of spectral technology can rapidly monitor the nitrogen nutrition of crops,which is helpful for managers to take timely measures to ensure crop yield,improve fertilizer efficiency and reduce environmental pollution.This thesis takes maize as the research object and combines pot experiment and field experiment.Using hyperspectral techniques to explore the temporal and spatial distribution of biochemical components and physical components of leaves and their correlation with spectral reflectance of leaves,leaf SPAD values and it's relationship between nitrogen nutrition,the relationship between leaf spectrum and plant nitrogen content(PNC)and the construction and evaluation of prediction models,the relationship between leaf spectra and root-zone soil inorganic nitrogen and the construction and evaluation of prediction models.The main conclusions are as follows:1.The biochemical components and physical indicators of the leaves exhibit certain spatial and temporal differences.Chlorophyll content,soluble protein content,thickness and net photosynthetic rate of leaves showed an increasing trend with the increase of nitrogen fertilizer application,showing a negative correlation with spectral reflectance in the visible band and a positive correlation in the near infrared band.Soluble sugar of leaves presented a decreasing trend with the increase of nitrogen fertilizer,showing a positive correlation with spectral reflectance in the visible band and a negative correlation with spectral reflectance in the near infrared band.2.With the passage of growth period,LNC presented a decreasing trend.In the same growth period,LNC at the same fertilization level: upper >middle > lower,LNC first increased and then stabilized with the increase of nitrogen fertilizer.The sensitivity of the lower leaves was stronger at the jointing stage,and the optimal ratio spectral index was RSI(1811,1842).The sensitivity of the upper leaves was stronger at the booting stage and the filling stage,and the optimal ratio spectral indexes were RSI(720,557)and RSI(600,511).The sensitivity of middle layer leaves was stronger at the anthesis-silking stage,and the optimal spectral index was RSI(688,644).3.The spatial and temporal distribution of SPAD value was determined.In the first three growth stages,the SPAD value of upper leaves had a stronger correlation with LNC,and the middle leaves had a stronger correlation with LNC at the filling stage.While the correlation between the SPAD value of the lower leaves and PNC reached the extremely significant level in all four growth stages,while the correlation between the average SPAD value of the leaves and PNC reached the extremely significant level.4.At the jointing stage,PNC was the largest,and showed a decreasing trend with the growth of maize.In the same growth period,with the increase of nitrogen fertilizer,PNC showed an increasing trend.The spectral reflectance of leaves at the booting stage and the filling stage was negatively correlated with the PNC in the visible band,while at the jointing stage and the anthesis-silking stage,there was a positive correlation in some near-infrared bands.5.Nitrogen application could increase significantly the content of root-zone soil inorganic nitrogen,and showed a decreasing trend with the growth of maize.The content of root-zone soil inorganic nitrogen at the jointing stage and the anthesis-silking stage was negatively correlated with the spectral reflectance of leaves in the visible band,positively correlated with the near-infrared band,and negatively correlated at the filling stage in the visible band.6.For the prediction of plant nitrogen content,the optimal fitting spectral indexes for the four growth periods were RSI(816,835),RSI(1030,738),PPR and RVI-2.PLSR models at the jointing stage,the booting stage and the filling stage could predict PNC well,while at the anthesis-silking stage could predict roughly.For the prediction of root-zone soil inorganic nitrogen content,the optimal fitting spectral indexes for the four growth periods were RVI-2,RVI-2,RSI(567,519)and RSI(802,816).Based on the prediction model established by PLSR,root-zone soil inorganic nitrogen content can be well predicted in the first three growth periods.PLSR model can be used to predict plant nitrogen content and soil inorganic nitrogen content in root zone by using leaf spectral data.