Study On Key Parameters Of Winter Rapeseed Growth And Yield Inversion By Remote Sensing

Rapeseed is one of the most significant oil crops in China,it plays an important role in the national economy.The advantages of in-time,rapid,accurate monitoring of rapeseed has an important and significant research meaning to production management and disaster prevention.Leaf nitrogen concentration?LNC?,above ground biomass?AGB?,nitrogen accumulation?NA?,growth period,leaf area index?LAI?and yield?Y?can accurately reflect the growth status and nutrition level of rapeseed.UAV remote sensing technology stands out in the field of crop growth monitoring due to its advantages of easy operation,saving time and energy,and low cost;the crop growth model has become an important method to study the growth of crops in recent years because of its comprehensive consideration of meteorological factors,soil factors,management parameters,and strong mechanism.In this paper,winter rapeseed with different nitrogen application levels and different growth stages were taken as the research objects,and rapeseed growth parameter prediction models based on empirical statistical models and crop growth models were established respectively.Taking the bud period as an example,the prediction results of the two methods were compared and analyzed,and the advantages and disadvantages of the two methods and the scope of application were discussed.The main conclusions drawn from the above research contents were as follows:Estimation of nitrogen nutrition parameters?LNC,AGB,NA?of winter rapeseed based on vegetation indices.Based on such as blue,green,red,red edge,and near-infrared bands,five-band images obtained by the drone,12 commonly used vegetation indices were calculated,and the correlations between these 12 vegetation indices with LNC,AGB,and NA were compared and analyzed.Taking the bud period as an example,four traditional regression models were analyzed using an empirical statistical model?linear function,exponential function,logarithmic function,polynomial function?,and it was found that the accuracy of the quadratic equation were better than the other equation?R²?,so we selected the quadratic equation as the best evaluation prediction model.Based on the quadratic function model,eight vegetation indices with higher R² and lower RMSE were selected for further sensitivity analysis?Noise Equivalent,NE?.The results showed that the red standard value NRI1 and the blue standard value NBI were best,because it's sensitive to LNC,AGB and NA and the estimation accuracy was high.The prediction determination coefficients R² of NRI1 for LNC,AGB and NA were 0.94,0.99 and 0.96 respectively,and the prediction determination coefficients R² of NBI for LNC,AGB and NA were 0.94,0.98and 0.98 respectively,which could estimate the nitrogen nutrition parameters of winter rapeseed.In the APSIM-Canola model,the main parameters affecting winter rapeseed growth include phenological parameters(CTT_JUV,CTT_FI,CTT_FL,CTT_{St GF},CTT_GF,VD_max,DL_min,DL_max)and biomass parameters?HI,RUE,Node phyllochron,leaf size,Node?no?app,leaf number?.Specifically,the change of each parameter will have an impact on the simulation results.You need to choose the appropriate parameters to adjust according to different needs when adjusting parameters.By adjusting parameters of the APSIM-Canola model,it was found that the APSIM-Canola model had a good prediction effect on the growth period,LAI,biomass and yield of winter rapeseed?the determination coefficient R² of the verification set was greater than 0.6?,and all points were evenly distributed on both sides of the diagonal in the measured and predicted diagram,indicating that the APSIM-Canola model could be used to predict winter rapeseed growth parameters,which provides a theoretical basis for winter rapeseed growth monitoring and yield estimation.Taking the bud period as an example,we compared the prediction effects of empirical statistical model and crop growth model on biomass.It was found that the empirical statistical model had a good prediction effect.In the 1:1 relationship between the predicted value and the measured value,all the points evenly distributed near the diagonal line,and the the R² of NRI1 was 0.86,the root mean square error RMSE was 587.47 kg/ha,the R²of NBI was 0.92,and the root mean square error RMSE was 647.52kg/ha.The predicted values of the APSIM-Canola model was slightly higher than the measured values,the determination coefficient R² was 0.61,and the root mean square error RMSE was801.93kg/ha.However,the APSIM-Canola model has the advantages of strong mechanism and good simulation effect on the actual growth environment.Therefore,in actual scenes,the model can be reasonably selected according to different needs.