GRACE时变重力场滤波方法  

Filtering method of GRACE time-variable gravity field

作　　者：李圳[1,2] 章传银[1] 柯宝贵[1] 

机构地区：[1]中国测绘科学研究院,北京100830 [2]山东科技大学测绘科学与工程学院,山东青岛266510

出　　处：《测绘科学》2017年第12期14-19,共6页Science of Surveying and Mapping

基　　金：国家重点研发计划项目(2016YFB0501702);国家自然科学基金项目(41374081;41574004)

摘　　要：针对GRACE时变重力场模型高阶项误差较大导致的"南-北"条带噪声,该文利用模拟的GRACE数据分析了去相关滤波、Gaussian滤波、组合滤波和平滑先验信息滤波方法对噪声的滤除效果和对真实信号的衰减程度。实验表明:4种滤波算法均能有效降低条带噪声,但单独使用去相关滤波时效果较差,需与其他算法结合使用;Guass滤波和组合滤波在减小噪声条带的同时,也在一定程度上牺牲了空间分辨率;平滑先验信息滤波在移除噪声、保留有效信号方面比其他3种算法有较为明显的优势。Aiming at the problem that time-variavle gravity field model solved by GRACE has greater error in its high-order,which is mainly expressed a kind of ＂south-north＂ band noise.This paper analyzed the effective of de-correlated filter,Gaussian filter,combined filter and SPM filter methods of noise removing and the attenuation of real signals using simulated GRACE data.The results revealed that the four filtering algorithms could effectively reduce the band noise,but using de-correlated filter in isolation is less effective,and it could be combined with other algorithms better;Guass filtering and combined filter could not only reduce the noise band,but also reduce the spatial resolution;spm filter has more obvious advantages in removing noise and retain effective signal comparing with the other three algorithms.

关 键 词：GRACE 时变重力场 滤波方法 

分 类 号：P223[天文地球—大地测量学与测量工程]