测绘通报 ›› 2022, Vol. 0 ›› Issue (2): 78-82.doi: 10.13474/j.cnki.11-2246.2022.0047

• 学术研究 • 上一篇    下一篇

联合LMD与EMD的GNSS站坐标时间序列去噪方法

熊常亮1,2, 贺小星3, 马下平4, 付杰5   

  1. 1. 中国科学院精密测量科学与技术创新研究院, 湖北 武汉 430077;
    2. 中国科学院大学, 北京 100049;
    3. 江西理工大学土木与测绘工程学院, 江西 赣州 341000;
    4. 西安科技大学测绘科学与技术学院, 陕西 西安 710054;
    5. 华东交通大学土木建筑学院, 江西 南昌 330013
  • 收稿日期:2021-02-28 修回日期:2021-05-21 发布日期:2022-03-11
  • 通讯作者: 贺小星。E-mail:hexiaoxing@whu.edu.cn
  • 作者简介:熊常亮(1998-),男,硕士生,主要从事GNSS数据处理与惯性导航研究。E-mail:xiongchangliang20@mails.ucas.ac.cn
  • 基金资助:
    国家自然科学基金(42061077;41904171);江西省自然科学基金(20202BAB21402);江西省教育厅科学技术研究项目(GJJ200639;GJJ171293);江西省水工安全工程技术研究中心开放研究基金(2020GGCZX04);江西省教育科学规划课题(21YB072)

Denoising method of GNSS station coordinate time series based on LMD and EMD

XIONG Changliang1,2, HE Xiaoxing3, MA Xiaping4, FU Jie5   

  1. 1. Innovation Academy for Precision Measurement Science and Technology, Chinese Academy of Sciences, Wuhan 430077, China;
    2. University of the Chinese Academy of Science, Beijing 100049, China;
    3. College of Architecture and Surveying Engineering, Jiangxi University of Science and Technology, Ganzhou 341000, China;
    4. College of Geomatics, Xi'an University of Science and Technology, Xi'an 710054, China;
    5. School of Civil Engineer and Architecture, East China Jiaotong University, Nanchang 330013, China
  • Received:2021-02-28 Revised:2021-05-21 Published:2022-03-11

摘要: 针对GNSS站坐标时间序列信噪不易分离的问题,在传统EMD去噪方法的基础上,本文提出了一种联合LMD与EMD的坐标时间序列去噪方法。该方法首先采用LMD分解原始坐标时间序列,基于连续均方误差(CMSE)原则分离高频噪声与低频信号,保持低频分量不变;然后对高频分量进行EMD去噪;最后以2次分解所得低频信号之和作为去噪后时间序列。以仿真数据与8个GNSS基准站实测数据进行试验,通过多种评价指标进行精度评估。结果表明,与传统EMD方法相比,联合LMD与EMD的方法能够更加精确地去除坐标时间序列中的噪声。

关键词: 全球卫星导航系统, 坐标时间序列, 局部均值分解, 经验模态分解, 去噪分析

Abstract: Aiming at the problem that the signal and noise of GNSS station coordinate time series is hard to separate, based on the traditional EMD denoising method, a new method of coordinate time series denoising combining LMD and EMD is proposed. This method first uses LMD to decompose the original coordinate time series, separates high-frequency noise and low-frequency signals based on the principle of consecutive mean square error(CMSE), keeps the low-frequency components unchanged, and then performs EMD denoising on the high-frequency components. Finally, the sum of the low-frequency signals obtained by the two decompositions is the time series after denoising. Experiments are carried out with simulation data and measured data from 8 GNSS reference stations, and the accuracy is evaluated through a variety of evaluation indicators. The results show that the combined LMD and EMD method can remove the noise in coordinate time series more accurately than the traditional EMD method.

Key words: global navigation satellite system, coordinate time series, local mean decomposition, empirical mode decomposition, denoising analysis

中图分类号: