TIME SERIES DENOISING BASED ON WAVELET DECOMPOSITION AND CROSS-CORRELATION BETWEEN THE RESIDUALS AND THE DENOISED SIGNAL

摘要

In this paper, a new denoising method, based on thernwavelet transform of the noisy signal, is described. Thernmethod implements a variable thresholding, whosernoptimal value is determined by analyzing the crosscorrelationrnbetween the denoised signal and the residualsrnand by applying different criteria depending on thernparticular decomposition level. The residuals are definedrnas the difference between the noisy signal and therndenoised signal. The procedure is suitable for denoisingrnsignals in real situations when the noiseless signal is notrnknown. The results, obtained with synthetic datarngenerated by well-known chaotic systems, show the veryrncompetitive performance of the proposed technique.