| Soil moisture is a key indicator in agricultural production,where soil moisture content in the root zone is related to the growth and development of crops at each stage,which in turn affects the final yield of crops.Accurate acquisition of soil moisture content in the root zone can provide technical support for water-saving irrigation and sustainable agricultural development.However,due to the significant differences in spatial distribution of soil water,the detection of soil water content in the root zone has obvious interference effects,and how to achieve the detection of soil water content in the root zone within a certain depth range is a current research challenge.To address this problem,this study used ground-penetrating radar to investigate the variation of dielectric constant with root density and soil water content in the presence and absence of roots by setting up indoor simulation experiments and outdoor field validation,and constructed a dielectric constant model applicable to root zone media.The inversion method of soil water content was further investigated,and finally it was compared with the AEA method and validated in the field in a large field.The main conclusions are as follows:(1)The relative permittivity of soil is closely related to the soil water content as well as the root content in the soil.At different root densities,the dielectric constant increases with increasing water content;similarly,at different soil water contents,the dielectric constant increases with root density.(2)The average amplitude envelope(AEA)was used to fit the average amplitude envelope cepstral and dielectric constant of the first half-cycle of radar waves,and the results of R^2were 0.844 and 0.913.The full waveform inversion method was used to construct root densitydielectric constant models at different water contents and root density-dielectric constant models at different water contents.The results of R^2 were greater than 0.9.Based on this,we constructed the soil water content and root density-dielectric constant models and concluded that the dielectric constant would increase with root density at the same water content.The dielectric constant gradually increases with the increase of root density at the same root density.(3)The application of the full-waveform inversion method can improve the inversion accuracy of soil water content in the root zone.In the indoor simulation experiment,the RMSE of soil water content inversion by AEA method was 0.018 and MAPE was 3.94%,while the RMSE of soil water content inversion by full waveform inversion method was 0.012 and MAPE was2.48%.Compared with the application of the AEA method,the RMSE of the full waveform inversion method was reduced by 0.006 and the MAPE by 1.46%.In the outdoor field experiment,the RMSE of the AEA method was 0.037 and the MAPE was 8.68%,while the RMSE of the full waveform inversion method was 0.028 and the MAPE was 5.58.Compared with the application of AEA method,the RMSE of full waveform inversion method was reduced by 0.009 and MAPE was reduced by 3.1%.The study shows that the use of the full waveform inversion method can effectively improve the accuracy of ground-penetrating radar in detecting soil water content in the root zone,providing technical support for accurate,rapid and continuous acquisition of soil water content at small and medium scale ranges in agricultural fields such as those containing root systems. |
