Extracting Characteristic Information Of Weak Signal From Strong Noise Background By Wavelet Analysis

The theory of Wavelet Transform is a great advance again after Fourier techniques in area of Mathematics. Because wavelet analysis has good characterics of time-frequency localization, it provide powerful tools for signal processing and resolve many problems which had not been done by Fourier Transform. The signal can be decomposed into different frequency components with the help of Muti-scale Analysis, and Wavelet methods offers an effective way for signal filtering, signal-noise separating, and characteric extracting. Therefore, in this thesis, Wavelet methods have been used to process the data of deformation monitoring for resolveing how to extract the characterics of weak signal disturbed by strong noise in the data of deformation monitering, the characteric of deformation signal and noise in wavelet transforming is dicussed, and deformation signal filtering is done with Muti-scale Analysis, the deformation signal is reconstructed with the wavelet decomposed coefficients processed by thresholding, and fantasticality point and its position of deformation monitored data is find out with fantasticality analysis. The superiority of wavelet analysis used in area of deformation data processing is proved by making use of four simulated examples and processing the data from GPS automatic surveying system to monitor Geheyan dam deformation, and some meaning advices are put forward.