高斯过程回归辅助下的GPS干涉反射积雪深度估测

GPS multipath effect snow depth estimation by Gaussian process regression-assisted

本文对全球定位系统干涉反射技术进行了研究。以美国板块边界天文台计划提供的P101测站的GPS监测数据为基础,利用GPS卫星高度角低于某一角度时多路径效应明显的特点,构建高斯过程回归(GPR)辅助的GPS干涉反射积雪深度估测模型,并监测了测站周围的积雪深度。结果表明,GPR辅助的GPS干涉反射积雪深度估测模型输出的雪深估测值的精度,相比传统单星反演结果有不同程度的提高,并且更贴近实测雪深的变化,为地表雪深反演提供了新思路。

This article studies the global positioning system-interferometric reflectometry technology.Based on the GPS monitoring data of P101 station provided by the Plate Boundary Observatory in the United States,it utilizes the obvious feature of multipath effect when the altitude angle of GPS satellite is lower than certain angle,constructs Gaussian process regression-assisted(GPR-assisted)GPS interference reflection snow depth estimation model and monitors the snow depth around GPS stations.The results show that the accuracy of the snow depth estimation value output by the GPR-assisted GPS interference reflection snow depth estimation model is improved to varying degrees compared with the traditional single-satellite inversion result,and the change trend of GPR snow depth estimation value is closer to the change of actual snow depth,which provides a new idea for surface snow depth inversion.