| In crop production,nitrogen(N)is an important limiting factor.Reasonable nitrogen fertilizer management is essential for efficient crop production and sustainable agricultural development,and can reduce the negative impact of nitrogen fertilizer increase.In this study,wheat was taken as the research object.By using portable canopy active sensors GreenSeeker and Crop Circle-ACS470,the growth and nutrient indicators monitoring models of wheat,optional vegetation index dynamic models and yield prediction methods based on optional vegetation indicators dynamic models were systematically studied.In addition,field experiments were undertaken to examine the performance of the models.These results not only provided a method for the non-destructive monitoring model based on canopy sensors of winter wheat,but also developed the optional dynamic vegetation index model.On that basis,the suitable vegetation index was extended in our study,which can provide the key technical support for yield prediction.Field experiments were conducted with different N application rates and wheat varieties.Firstly,wheat growth and nutrient monitoring models were studied.Five spectral reflectance values were obtained by the active canopy sensor of Crop Circle ACS-470,fifty-six vegetation indices were established based on these reflectances.In these fifty-six vegetation index,we selected 10 best-performing indices,established the winter wheat growth and nutrient monitoring models with these 10 indices and the index NDVI,RVI obtained from GreenSeeker,by evaluating and comparing the correlation coefficient(R2),relative root mean square error(RRMSE)and noise equivalent(NE),it is found that the optimal vegetation index RERVI and NDRE obtained from Crop Circle ACS-470 performed better than NDVI and RVI obtained from GreenSeeker.Based on the monitoring models of winter wheat growth and nutrient indicators,the optimal vegetation index RERVI was selected by comparison,and the dynamic model based on optimal vegetation index RERVI was further studied.Crop Circle-ACS470,as a portable active canopy sensor,objectively reflects the growth of canopy,and the canopy refers to the leaf area index LAI.Therefore,we analyze the relationship between LAI and exponential RERVI,and analyze the relationship between LAI and RERVI.Based on the monitoring models of winter wheat growth and nutrient indices,the optimal vegetation index RERVI was selected by comparison,and the dynamic model based on optimal vegetation index RERVI was further studied.Crop Circle-AC S470 as a portable active canopy sensor,objective to reflect the growing crop canopy,canopy refers to the leaf area index LAI.So,we analyzed the relationship between LAI and index RERVI,and analyzed the relationship between LAI and RERVI from the mechanism.We found that RERVI and LAI relationship is correlation,and have the same trend.After the data were analyzed,we established the Gauss model by RERVI dynamic change with AGDD at different yield levels,divided into low yield level(yield<4.7 t/ha),middle level(4.7 t/ha<yield<7.0 t/ha),high yield level(yield>7.0t/ha).the dynamic changed of wheat index based on different yield levels were studied.The three production gradients were distinguished.The model validation results are good and feasible.Based on the studied of the dynamic model,this study attempts to combine the continuous uninterrupted exponential RERVI integral J RERVI with the yield to construct a yield prediction method.This study uses continuous RERVI,ie,integral ?RERVI,for a single fertility period to predict the yield and avoid the error of a single measurement.The results show that the accuracy of ?RERVI prediction is low before jointing,and the accuracy is improved after jointing.With the growth period,the accuracy is getting higher and higher,and the accuracy reaches its peak until the filling stage.The yield prediction method based on the optimal vegetation index RERVI dynamic model can provide a new reference for crop grain yield prediction. |
PDF Full Text Request