Denoising Algorithm Of Wavelet Modulus Maximum Based On Blackman Window Interpolation

Research of signal denoising algorithm is the difficulty and hot spot in the signal processing,however, Fourier transform has long used as the major methods to analyze and process the signals and the result is good. Fourier transfor m is a whole domain transformation, which displays the characterist ic of signal only in a dom ain, so Fourier transform showed with great limitation to non-stationary signals.The wavelet transform is a forceful m eans which has been devel oped rapidly to deal with non-stationary signals after the Fourier transform, it possesses the features of localization in tim e and frequency domain at the same time and with the ability of multi-resolution analysis. Not only has wonderfully potential application at signal processing, and formed a new force by mutual penetration and fusion with other fields.This paper mainly studies the signal reco nstruction algorithm from reconstruction wavelet coefficients by wavelet modulus maximum. The m odulus maxima direct reconstruction algorithm and th e alternating projection algo rithm are analyzed and the advantages and disadvantages of the tw o algorithms are summ arized. Found that modulus maxima direct rec onstruction algorithm is si mple and efficient, but the incomplete information of the reconstructed signal and larger distortion; the alternating projection algorithm has a high precision, but the large amount of calculation, slow operation speed and poor practicability. Ho w to find a reasonable and effective, convenient and quick algorithm to reconstruct signal from remained signal m odulus maxima is a need to be solved problem.Considering the Blackman window function ha s the characteristics of wide m ain lobe and larger am plitude, small side-lobe amplitude and fast attenuation, we use the Blackman window function as the interpolatio n function. On this basis, a denoising algorithm of wavelet modulus maxim um based on Blackm an window function interpolation is structured. This algorithm is using the interpol ation technology to fill the points which are not m odulus maxima on each scale and obtain th e reconstructed signal. The simulation is made on the reconstruction precision, SNR gain and operation time to the blocks and bum ps signal by co mpared with two class ic algorithms. The results show that d enoising algorithm based on Blackm an window function interpolation has a better denoised effect w ith the complete signal information, small amount of calculation, small distortion, quick convergence and stronger practicality.