Crop Growth Monitoring Based On Polarimetric Radar Remote Sensing

In the field of agriculture,crop growth monitoring is critical to the safety of grain markets.Crop lodging is one of the common agricultural disasters.Plant water content and crop biomass are important indicators in the process of crop growth.It is of great economic significance to monitor crop disasters and growth.Remote sensing technology provides a means for large-scale crop growth monitoring.Traditional optical remote sensing,due to the influence of bad weather such as fog,can't effectively get crop information on the ground,which is not conducive to real-time monitoring of crop growth.Synthetic Aperture Radar(SAR)has strong penetration and is not affected by weather conditions,which makes SAR as a long-term,multi-environment agricultural monitoring means.In view of the above background,this study takes maize and wheat as research objects,carries out satellite-to-ground synchronous experiments,and conducts dual-polarized SAR and full-polarized SAR research on Maize lodging monitoring,wheat plant water content inversion and wheat aboveground biomass inversion.The following work has been completed:The study simultaneously acquired the experimental areas(Beijing Xiaotangshan National Precision Agriculture Research Demonstration Base and Gaocheng District,Shijiazhuang City)dual-polarized Sentinel-1A radar data and ground measured data of maize lodging and wheat growth.The study built and analyzed the relevant radar polarized index,and established the quantitative monitoring model of maize lodging and water content of the wheat plant were retrieved by water cloud model under regional scale.The empirical model and semi-empirical model were used to quantitatively analyze the monitoring results of radar polarized characteristics on maize lodging and wheat plant water content;The study simultaneously acquired the experimental areas(Shangkuli farm,Inner Mongolia Autonomous Region and Gaocheng District,Shijiazhuang City)full-polarized GF-3 radar data,Sentinel-1A/B radar data and ground measured data of wheat growth.The study extracted GF-3 radar polarization decomposed information,and a polarized water cloud model(APWCM)was proposed for wheat biomass inversion.Comparative analysis wheat biomass inversion results of GF-3 radar data and Sentinel-1 radar data in two study areas based on water cloud model and APWCM.The study used a mechanism model to quantitatively analyze the monitoring results of radar polarized characteristics on wheat biomass.The study got the following conclusions(1)The most sensitive backscattering coefficients of maize plant height before and after lodging are ?0VH and ?0VV+VH,respectively.The R2 of the measured and simulated difference of maize plant height in the sample points of the modeled set is 0.896.The overall correlation between the quantitative ratio result estimated by the model and the measured ratio is 0.899 The correlation of moderate lodging was the best,followed by severe lodging and mild lodging.The accuracy of the lodging degree classified results of the verified set sample points and the total sample points is 100%.The results provide a reference for quantitative monitoring of maize lodging based on dual-polarized Sentinel-1 radar data and effective grading monitoring of maize lodging(2)Based on the dual-polarized Sentinel-1A radar data,the modeled accuracy of the wheat plants water content inversion model is:RMSE=0.022,nRMSE=19.98%,and the validated accuracy of model is RMSE=0.026,nRMSE=21.24%.The study provides a methodological reference for monitoring the water content of wheat plants based on Sentinel-1 radar data(3)The wheat biomass inversion results based on full-polarized GF-3 radar data and APWCM in the two study areas are as follows:RMSE=210.85g m-2,nRMSE=17.91%;RMSE=743.58 m-2,nRMSE=13.53%.The inversion results of wheat biomass based on water cloud model in two research areas generally have smaller errors.The accuracy of wheat biomass inversion results based on GF-3 radar data and water cloud model was the highest in the two study areas(RMSE=131.63g m-2,nRMSE=11.18%;RMSE=645.17g m-2,nRMSE=11.74%).However,APWCM,as a physical model,does not need additional ground-based auxiliary data for wheat biomass inversion,and has lower relative error in the areas with larger vegetable coverage.Therefore,the model can be used as an alternative method for wheat biomass inversion.