GNSS-IR Retrieval of Soil Moisture in Sugarcane Plantation Based on Cross-Correlation Satellite Selection Method

Timely and accurate monitoring of soil moisture in farmland is of great significance to the evaluation of crop growth and drought. Global Navigation Satellite System interferometric reflectometry (GNSS-IR), as a new remote sensing technology, can invert soil moisture based on the signal-to-noise ratio recorded by the measuring receiver. At present, existing studies tend to invert soil moisture in bare soil or low vegetation cover environments, and satellite selection depends on empirical values or prior information. Accordingly, a multisatellite combination method based on cross-correlation satellite selection for soil moisture inversion is proposed. Firstly, the trend and modulation terms in the signal-to-noise ratio of each satellite are effectively extracted by wavelet analysis. The characteristic of the wave term is analyzed, and the arc segment with an obvious and stable periodic oscillation is selected. Then, based on the interference phase of each satellite obtained by nonlinear least squares fitting, a cross-correlation satellite selection method (CCSSM) is established. The available satellites are selected by setting a reasonable threshold. Finally, three multi-satellite combination models for soil moisture inversion are constructed, and the inversion effects of each model are compared and analyzed. Taking the sugarcane planting area as an example, the results indicate that satellites can be screened quickly and effectively by CCSSM, and the selected satellites have strong cross-correlation. For short-term Global Navigation Satellite System (GNSS) observation data, it is more advantageous to use multiple linear regression model to invert soil moisture than machine learning.