Sensitivity Analysis And Hyperspectral Diagnosis Of Nitrogen In Summer Maize Leaves Under Different Nitrogen Application Conditions

Objective:The real-time and accurate monitoring of crop nitrogen nutrition is the key to realize the scientific and rational application of nitrogen fertilizer,and is also an important measure to improve the utilization rate of nitrogen fertilizer,protect the ecological environment and realize clean production.In this study,hyperspectral remote sensing technology was used to explore the rapid,accurate and non-destructive monitoring of nitrogen nutrition in summer maize.Methods:The distribution characteristics of nitrogen nutrients in the canopy at different growth stages were analyzed through single factor fertilizer effect tests with different nitrogen gradients,and hyperspectral information extraction techniques and methods were explored to establish an accurate prediction model of nitrogen content in leaves,providing theoretical guidance for nitrogen fertilizer application in summer maize.Results:1.With the increase of nitrogen application,the yield of summer maize showed a trend of first increasing and then decreasing.The analysis of yield component factors showed that nitrogen fertilizer had little influence on panicle number,and the hundred grain weight and panicle number were closely related to the increase of yield.2.In the leaf-spreading stage of summer maize,there were significant differences in nitrogen content between leaf positions D1 and other leaf positions.There was significant difference in nitrogen content in D10 leaves during leaf spreading period,which was relatively sensitive to nitrogen nutrition.There was significant difference in nitrogen content in leaves of D12 at spinning stage,which was sensitive to nitrogen deficiency.In the middle stage of grouting,D1 leaf position is relatively sensitive to nitrogen nutrition.With the decrease of leaf height,the nitrogen content of leaves showed a trend of first increasing and then decreasing,and the four key growth stages showed similar changes.3.Correlation analysis was conducted between vegetation index,newly constructed ratio vegetation index and nitrogen content in leaves.Stepwise regression and partial least square method were used for in-depth analysis of nitrogen content and hyperspectral information in leaves to construct a nitrogen content prediction model in leaves.Accurate nitrogen monitoring modelswereestablishedforleavesatthreegrowthstages:Y=150.29RSI(677,676)-148.98(R~2=0.84)at leaf spreading stage for 8,Y=88.87RSI(720,719)-89.35(R~2=0.91*)at leaf spreading stage for 16,Y=-2.303RSI(692,394)+5.379(R~2=0.91*)at middle grouting stage.The stepstep regression model Y=14.428RSI(780,740)-14.023(R~2=0.92*)at the spinning stage could accurately predict the nitrogen content of the leaves.Conclusion:A diagnosis model of nitrogen content in leaves of summer maize in four growth stages was established to provide theoretical guidance for rapid and accurate monitoring of nitrogen content in leaves.