摘要
为降低噪声对GNSS坐标时间序列的影响、有效提取时间序列中的有用信息,在局部均值分解(LMD)降噪方法的基础上引入奇异值分解(SVD)方法,建立了LMD-SVD方法。首先通过LMD方法将时间序列分解为若干个乘积函数(PF)和余量,PF分量可反映时间序列的时频分布特性;然后通过连续均方根误差方法确定高频分量与低频分量的分界点;最后对经SVD方法降噪后的高频分量、低频分量和余量进行重构,得到最终降噪结果。利用5个GNSS测站U方向坐标时间序列对该方法进行验证。结果表明,相较于单一LMD方法,LMD-SVD方法结果的信噪比与相关系数分别提高了34.28%与17.11%,均方根误差降低了51.31%,降噪效果更好。
In order to reduce the impact of noise on GNSS coordinate time series and effectively extract useful information in time series,based on local mean decomposition(LMD)denoising method,we introduced singular value decomposition(SVD)method to establish a LMD-SVD combined denoising method.Firstly,we decomposed the time series into several production functions(PF)and residual by LMD method,and PF could reflect the time-frequency distribution characteristics of time series.Then,we determined the dividing point between high-frequency compo-nent and low-frequency component by continuous root mean square error method.Finally,we reconstructed the high-frequency component,low-frequency component and residual after denoising by SVD method to obtain the final denoising result.We used the U-direction coordinate time series of five GNSS stations to verify this method.The results show that compared with the single LMD method,the signal-to-noise ratio and correlation coefficient of results of this method are increased by 34.28%and 17.11%respectively,the root mean square error is reduced by 51.31%,and the noise reduction effect is better.
作者
龚旭峥
汪香梅
王凯时
GONG Xuzheng;WANG Xiangmei;WANG Kaishi(Zhejiang Institute of Surveying and Mapping Science and Technology,Hangzhou 310023,China;Kaihua County Resource Planning Data Center,Quzhou 324300,China)
出处
《地理空间信息》
2024年第3期43-46,共4页
Geospatial Information
基金
浙江省自然资源厅2022年度科技资助项目(2022-54)。
