Nonlinear Wavelet Denoising of Data Signals

The purpose of this work is to develop a new strategy for filtering transmitted (or received) noisy data signals with minimum loss of information, which can be implemented efficiently. This strategy consists of a wavelet domain filtering that combines a discrete wavelet transform based decomposition phase and a wavelet coefficients shrinkage phase. Data-like signals are generated, white noise is added, the results obtained after filtering are evaluated through signal to noise ratios and performed gains. The proposed nonlinear denoising procedure performs a decomposition of identified noise and tries to find correlation between the data signal and noise. This correlation will be removed from the noisy signal. The denoising method is carried out in Matlab environment with simulated signals and added gaussian white noise. The results for different filtering techniques and different wavelet decompositons are compared through signal to noise ratio and obtained gain.