关键词: 局部均值分解, 奇异值分解, 连续均方误差, 奇异值差分谱, 坐标时间序列

Abstract: In order to effectively extract useful information from coordinate time series of GNSS station, and reduce noise interference, this paper proposes a signal denoising method that combining local mean decomposition with singular value decomposition. Experiments were carried out using measured coordinate time series of five stations. Firstly, the coordinate time series is decomposed into a series of PF components and residuals by local mean decomposition, and then the continuous mean square error method is used to determine the boundary between the high frequency component and the low frequency component. Keep the low-frequency components unchanged, and use the singular value decomposition method to denoise and reconstruct the high-frequency components. Finally, the reconstructed high-frequency components and low-frequency components are superimposed to obtain the final de-noising coordinate time series, and the noise reduction effect is compared and analyzed. The results show that compared with the simple singular value decomposition, the local mean decomposition combined with singular value decomposition can adaptively select the appropriate number of singular values for signal reconstruction, which improves the noise reduction effect.

Key words: local mean decomposition, singular value decomposition, continuous mean square error, singular value difference spectrum, coordinate time series