基于STL的GNSS高程时间序列分析  

Analysis of GNSS elevation time series based on STL

作　　者：李辰风 田坤 陈文 LI Chenfeng;TIAN Kun;CHEN Wen(Three Gorges Geotechnical Engineering Co.,Ltd.,(Wuhan),Wuhan 430074,China;Institute of Territorial Spatial Survey and Planning of Jiangxi Province,Nanchang 330025,China)

机构地区：[1]长江三峡勘测研究院有限公司(武汉),湖北武汉430074 [2]江西省国土空间调查规划研究院,江西南昌330025

出　　处：《水利水电快报》2025年第3期52-58,共7页Express Water Resources & Hydropower Information

基　　金：长江三峡勘测研究院有限公司(武汉)科研项目(SXYKY-2023-08)。

摘　　要：为了解决大规模GNSS坐标序列季节性信号和长期趋势项分离时的效率问题,引入基于局部加权回归的季节项趋势分解(STL)方法,分析该方法在GNSS时间序列分析中的应用效果。通过比较分析,利用STL分离提取模拟信号和GNSS高程时间序列的季节项和长期趋势项。结果表明:STL方法能够准确地分离出GNSS高程时间序列中的季节性信号和长期趋势项,与其他常用时间序列分析方法相比,该方法运算效率较高。STL方法的高效性和精确性使其非常适合应用于大规模GNSS坐标序列的信号分离。To address the efficiency issues in separating seasonal signals and long-term trend components from large-scale GNSS coordinate time series,we introduced the seasonal-trend decomposition using Loess(STL)method based on local weighted regression,and analyzed its application effectiveness in GNSS time series analysis.Based on comparative analysis,STL was applied to separate and extract seasonal and long-term trend components from simulated signals and GNSS elevation time series.The results showed that the STL method can accurately separate the seasonal signals and long-term trend components in GNSS elevation time series.Compared to other commonly used time series analysis methods,this method had a higher computational efficiency.STL is very suitable for signal separation in large-scale GNSS coordinate time series due to its efficiency and accuracy.

关 键 词：卫星大地测量 STL GNSS 时间序列 季节性信号 

分 类 号：P228[天文地球—大地测量学与测量工程]