| Soil moisture is an important physical quantity in the research of hydrology,climate and agriculture.It is an important determinant of global climate and ecological environment.GNSS-R(GNSS Reflectometry)technology has the advantages of no transmitter,a large number of signal sources,strong signal penetration ability,low cost and high temporal and spatial resolution.It has become a hot research direction of soil moisture monitoring.The research on Soil Moisture Retrieval in China,especially on airborne platform,started late,and most of them are based on foreign data.In this thesis,four problems are studied: the advantages and disadvantages of different inversion models for airborne soil moisture inversion,the inversion accuracy under different conditions,the improvement effect of the accuracy of airborne soil moisture inversion results by using correction methods such as soil roughness and antenna gain,and the system optimization of dual antenna receiver for Airborne GNSS-R soil moisture inversion.The main work and research contents are as follows:(1)The general situation of GNSS(Global Navigation Satellite System)system and soil moisture measurement is introduced.The types,geometric relationship,correlation power model and polarization characteristics of GNSS reflected signals are analyzed in detail.The key steps in the process of airborne soil moisture inversion are introduced and summarized in detail,mainly including the calculation of reflectance,various correction methods of reflectance,and the principles,advantages and disadvantages of four common soil moisture inversion models: Wang,Topp,hallikainen and Dobson.(2)Based on the empirical / semi empirical model,the UAV GNSS-R inversion experiment is carried out to verify the applicability and inversion accuracy of each inversion model.When the experimental data are relatively stable,the inversion results of Wang model are closest to the measured data,the inversion results of Dobson model are the most different from the measured values,and the inversion results of Topp model and hallikainen model are very close,which is between Wang model and Dobson model,and the inversion results of each inversion model are less than the measured values.In the experiment on December 20,the comprehensive inversion accuracy of PRN 29 satellite is the highest,with an average value of 0.1875cm3/cm3,which is very close to the measured value of 0.185 cm3/cm3.The standard deviation is 0.057,RMSE is 0.0574,MAPE is 0.3128,and the inversion accuracy of PRN 26 is the second.The inversion results of PRN 27 satellite have the smallest variance and the most stable.The inversion results of PRN 4 and PRN 16 satellites with high elevation are also relatively stable.The experiments verify the advantages of high elevation satellite in airborne GNSS-R soil moisture inversion.It is found that the inversion results are greatly affected by the instrument stability,site environmental factors and the length of acquisition time.Good experimental environment and stable long time series data are of great help to improve the accuracy.(3)The effects of surface roughness correction and antenna gain correction on the accuracy of airborne GNSS-R soil moisture inversion results are studied to verify the inversion accuracy when there is a large difference in soil moisture before and after rain.The inversion results show that after the roughness correction,the standard deviation increases by 5.6%,RMSE decreases by 21.2%,Mae and MAPE decrease by37.1%;After the experiment after rain,the standard deviation increases by 12.5%,RMSE decreases by 12%,Mae and MAPE decrease by 11.6%.Therefore,the roughness correction has an obvious effect on improving the accuracy of inversion results.Compared with the standard deviation before rain,the inversion accuracy after rain increases by 33.2%,RMSE by 72.6%,MAE by 109.5% and MAPE by 18.1%.Wang model is more sensitive to roughness correction,with the largest increase,Dobson model has the smallest increase,Topp model and hallikainen model have the middle increase and little difference.The result of multi model fusion is used as the result of roughness correction,and the result is relatively stable.An antenna gain correction method considering antenna directivity is proposed,and two correction functions suitable for airborne and ground-based platform antenna field of view angle of 0 ° and not 0 ° are given.(4)The dual antenna receiver system is improved.For the soil texture of Xuzhou,the top model and hallikainen model are added to the inversion module of the earth observation system.The airborne GNSS-R data processing software is developed based on MATLAB app dsigner platform.The software can post process the original dat data of airborne or ground-based experiments of dual antenna earth observation system,visually draw the key steps of processing,save the results and generate accuracy analysis report. |
PDF Full Text Request