| As an indicator of surface dryness and wetness,soil moisture is one of the key parameters in the global water cycle process,playing a very important role in agricultural production,meteorological research,and disaster warning.Therefore,it is of high scientific and practical value to study how to obtain soil moisture information with high efficiency,high precision and long period.In recent years,with the development of satellite remote sensing science and technology,the use of Global Navigation Satellite System Reflection Signal GNSS-R(Global Navigation Satellite System Reflectometry)for ground remote sensing has become an emerging microwave remote sensing monitoring technology.GNSS-IR(Global Navigation Satellite System Interference and Reflectometry)based on multipath effect has become a research hotspot of soil moisture inversion technology because of its advantages of low cost,no damage to observation objects,abundant signal sources,high resolution,and long-term continuous observation.This article proposes a signal fusion GNSS-IR soil moisture inversion method based on wavelet analysis,and the experimental data is sourced from the self-built GNSS-IR observation station.In order to weaken the error caused by signal jumps,the trend items were removed by db4,coif5,sym5,and mexh wavelet analysis respectively.Then,the entropy fusion method,adaptive fusion method and mean fusion method are respectively used to fuse the satellite signals to further improve the inversion effect of soil moisture.The main research content and results are as follows:(1)The GNSS observation data of the 40 th to 145 th annual product days of the self-built station in 2022 were selected as experimental data,and the measured soil moisture data was used as the reference data.By comparing the frequency,amplitude,and phase feature parameters extracted from the original Signal Noise Ratio(SNR)data,it was found that the amplitude has the highest correlation with the measured soil moisture.(2)The wavelet analysis method of coif5,sym5,db4,mexh and second-order polynomial method were used to remove trend items from the original Signal Noise Ratio(SNR)data.The SNR sequence characteristic parameters after removing the trend term were extracted and compared with the measured soil moisture,which proved that the wavelet analysis method could effectively improve the signal fitting accuracy.By analyzing the accuracy indexes of S2 X and S2 P signals of GPS system SNR,it can be found that the inversion accuracy from high to low are db4,coif5,sym5,mexh,and second-order polynomial method.Moreover,compared with the inversion results of second-order polynomial method,the Correlation Coefficient(R)of the two signals after db4 wavelet analysis increased by 13.21% and 11.78% respectively,the Root Mean Square Error(RMSE)decreased by 7.82% and 10.33%,respectively and the Mean Absolute Error(MAE)decreased by 17.24% and 10.05% respectively.(3)In order to make full use of the difference and complementarity between different satellites and different types of SNR signals,entropy fusion method,adaptive fusion method and mean fusion method were applied to the amplitude parameters of GPS satellite S1 C,S2X,S2 P and S5 I signals respectively on the basis of db4 wavelet analysis,and soil moisture inversion and modeling analysis were carried out.The results show that compared with the inversion effect of singlesatellite and single-frequency before fusion,the overall inversion effect of multi-satellite and multifrequency signal fusion is significantly improved,the R increased by 11.75%-43.62%,the RMSE decreased by 32.06%-66.25%,and the MAE decreased by 23.44%-71.75%.Moreover,the inversion ability of the three signal fusion methods is adaptive fusion,mean fusion and entropy fusion in order from high to low.The above studies proves that the combination of wavelet analysis and signal fusion method can effectively improve the stability and accuracy of GNSS-IR soil moisture inversion,which provides a reference for the application of this technology in practice. |
PDF Full Text Request