小波分析在垂向海潮特征信息去噪研究

A study of de-noising of vertical tide characteristic data based on wavelet analysis of the

目前国内外许多学者利用沿海GPS站点的高精度坐标时间序列进行海潮特征信息的提取研究,已取得一定成果,但高精度坐标时间序列中不可避免地存在许多噪声,极大影响海潮特征信息的提取。基于日本GMSD站的67d的高精度PPP坐标时间序列,利用小波分析进行去噪实验,经FFT变换后,将提取的结果与FES2004海潮模型的特征值比较。实验结果表明:经小波分析后,GPS站点的时间序列精度得到提高;经小波去噪后的反演精度均有不同程度的提高,最多达到0.14mm,而海潮特征值的量级为厘米级或者亚厘米级,说明小波分析对海潮特征值的提取精度有明显提高。

Currently many scholars at home and abroad use high precision coordinate time series in coastal GPS stations to extract the characteristic information of the tide, which has made many achievements. But high precision coordinate time series contain lots of inevitable noise, which affect the extraction in the characteristic information. In this paper, based on 67 days coordinate time series of the result of precise point positioning （PPP） of the Japanese GPS station （GMSD）, wavelet analysis is used to make de-noising experiments. After Fast Fourier Transform （FFT） method, by comparing the extraction with FES2004 tidal model of the eigenvalue. The results show as follows, the wavelet analysis can improve the aceuracy of coordinate time series; wavelet analysis can improve the aceuracy of eigenvalue, whieh is up to 0. 14mm. Compared with scale in mm- cm level between eigenvalue of the tide, the accuracy of wavelet analysis in inversion precision is quite required.