Wavelet Correlation Analysis of Geodetic Signals

Classical correlation reflects linear relation between two signals completely in frequency domain, while wavelet correlation, introducing scale parameter to the classical one, is studied to realize similarity degree analysis both in time domain and frequency domain. So the latter is superior to analyzing correlation between two non-stationary geodetic signals. Wavelet correlation can detect similarity degree at different frequencies and delays and give expression to delay information while the correlation obtains maximum at a certain frequency. In order to determine the scales in which feature information lies, wavelet spectrum, combining wavelet transform and Fourier spectrum analysis, is used to explore feature information at different scales. It can be seen cycle information- cycles of a month, a season, half of a year, and a year-hidden in signals from stations in Shandong through wavelet spectrum analysis. Then linear relation between signals from two stations at three directions, North, East and Up, is analyzed by wavelet correlation respectively.