Study On Nitrogen Detection And Zoning Of Soybean Canopy Leaves Based On Multi-spectral UA

Making variable fertilization decision based on farmland canopy information is an effective means to control fertilizer waste and improve the utilization rate of chemical fertilizer.The UAV equipped with a multi-spectral camera can realize the efficient and accurate collection of crop canopy information.For canopy leaf nitrogen distribution of soybean in different growth period,based on the principle of nitrogen spectral detection and leaf reflection under different bands,soybean as the research object,the establishment of soybean under different varieties and different fertility canopy nitrogen detection model,and based on nitrogen distribution reasonable partition,for soybean canopy nitrogen information fast acquisition and key topdressing period variable fertilization management lay the theoretical basis.Focus on the following aspects of the work research.To obtain the spectral variable data of longken 3401,Heihe 43,longken 310(soybean varieties)in early flowering(R1),early pod(R3),primary grain(R5)(soybean growth cycle),and measured the total nitrogen data of the study block canopy leaves by chemical method,analyzed the correlation between spectral variables and nitrogen,and established the linear,binomial,power and index function model of nitrogen,respectively.Meanwhile,the best spectral variable combination was selected by multiple linear regression modeling method,and the canopy nitrogen detection model of soybean leaves was established through comparative analysis.It is concluded that the determination coefficient of the regression model is above 0.83 and the RMSE error is below 0.2.The results show that the built model can provide a reference basis for the detection of nitrogen elements in soybean canopy leaves.Using the established model for the experiment area of nitrogen inversion,through K-Means cluster analysis and the elbow inflection point of nitrogen reasonable classification,the same soybean field distribution and the actual growth situation,further verify the reliability of soybean canopy nitrogen diagnostic model,and provide theoretical support for the next variable fertilization work.