基于连续小波变换及正则化方法的潮位分析

Tidal Level Analysis Based on Continuous Wavelet Transform and Regularization Method

潮位的精确检测对于海洋监测、环流分析等是非常重要的课题,更是我国海洋事业发展至关重要的一环.以往的研究表明:现成的全球定位系统(GPS)可以利用多路径干涉的特性来监测潮位.Lomb-Scargle谱分析(Lomb-Scargle Periodogram,LSP)通常用于从信噪比(Signal-to-Noise Ratio,SNR)序列中提取多路径频率,为了提高其数据分辨率,可以使用连续小波变换来提取SNR序列的瞬时频率,并据此获得瞬时潮位.由于瞬时潮位数据噪音很大,因此存在着许多异常值和误差.通过分析多频多模站点数据,同时利用同卫星不同类型的SNR数据,在保连续性的基础上进行一定的正则化降噪处理.数值结果表明,正则化方法可以有效降低噪音,并提高潮位分析精度.

Accurate detection of tide level is a very important topic for ocean monitoring,circulation analysis,and so on,and it is also a crucial link for the development of China's marine industry.Previous studies have shown that the ready-made global positioning system(GPS)can use the characteristics of multipath interference to monitor the tide level.Lomb scalar periodogram(LSP)is usually used to extract multipath frequency from signal-to-noise ratio(SNR)sequences.In order to improve the resolution of the data,continuous wavelet transform can be used to extract the instantaneous frequency of the SNR sequence and obtain the instantaneous tide level.Based on this,the instantaneous tidal level is obtained.Because the instantaneous tide level data is very noisy,there are many outliers and errors due to the noise of the tide data.This paper tries to use different types of SNR data of the same satellite from multi-frequency and multi-mode stations at the same time,and perform a certain regularization noise reduction processing on the basis of preserving continuity.The numerical results show that the regularization method can effectively reduce noise level and improve the accuracy of tide level analysis.