An Improvement Algorithm Based on Wavelet Shrinkage De-noising

摘要

Wavelet threshold processing is an effective method of signal de-noising. Based on researching about the characteristics of traditional wavelet de-noising algorithm, this paper proposes a new transformation function with transforming wavelet coefficients, which can magnify the areas where signals and noises are easily confused. Next, an effective threshold and a new adaptive threshold function are structured to filter noise signals in wavelet decomposition scale and wide scale by combining the advantage of the hard threshold and the soft threshold functions. The simulation results on blocks and bumps signals demonstrate that Signal-to-noise Ratio (SNR) and Root-mean-square Error (RMSE) from the improved algorithm are much better than traditional threshold filtering method. In addition, the new method can get rid of noises and retain useful feature information more efficiently.