基于GPS-IR的土壤湿度多星非线性回归估算模型

Multi-star Nonlinear Regression Estimation Model of Soil Moisture Based on GPS-IR

由于传统方法利用单星进行土壤湿度估算精度较低,且干涉相位与土壤湿度之间的线性关系与实际情况存在差异,GPS-IR (global positioning system-interferometry reflectometry)技术可用于反演土壤湿度。考虑到多星联合反演土壤湿度具有优势,提出一种多星非线性回归估算模型。实验结果表明,该模型有效改善了单星反演出现异常跳变的现象;采用双星或三星联合建立的多星非线性回归模型的估算效果要优于单星,反演结果与土壤湿度参考值之间的相关系数r分别达到0.90和0.93以上,RMSE平均值分别达到了0.029和0.026。

Because soil moisture estimated by traditional methods has low accuracy,and the linear relationship between the interference phase and soil moisture is different from the actual situation,GPS-IR technology can be used to invert soil moisture.Considering the advantages of multi-star joint inversion of soil moisture,this paper proposes a multi-star nonlinear regression estimation model.The experimental results show that:the model effectively improves the phenomenon of abnormal jumps in single star inversion;the multi-star nonlinear regression model established by the combination of a double star or threestar is better than that established by a single star,the correlation coefficient r between the inversion result and the soil moisture reference value is 0.90 and 0.93,respectively,and the RMSE averages reached 0.029 and 0.026,respectively.