Signal De-noising method based on particle swarm algorithm and Wavelet transform

Wavelet analiza je novi alat za analizu odnosa vrijeme-frekvencija, razvijen na temelju Fourierove analize s dobrim svojstvom lokaliziranja vremena i frekvencije i mogućnosti donošenja višestrukih rješenja. Koristi se u cijelom nizu primjena u području obrade signala. U ovom se radu analizira primjena wavelet transforma u filtriranju signala korištenjem poboljšane optimalizacije roja čestica i predlaže inteligentna metoda uklanjanja šuma iz signala zasnovana na wavelet analizi. Metoda koristi Center Based Particle Swarm Algorithm (CBPSO) za izbor optimalnog praga za svaki pod-pojas u različitim mjerilima, inteligentno razaznavajući vrstu šuma iz samog signala, što ne zahtijeva nikakvo prethodno poznavanje šuma. Poboljšani algoritam roja čestica koristi se da potakne optimalni izbor različitih mjerila praga wavelet domena, što je dovelo do uklanjanja šuma iz signala kod različitih tipova pozadinskog šuma, i povećane brzine wavelet transforma i wavelet konstrukcije te ima veću fleksibilnost. Eksperimentalni rezultati su pokazali da se CBPSO algoritmom može postići bolji učinak uklanjanja šuma.Wavelet analysis is a new time-frequency analysis tool developed on the basis of Fourier analysis with good time-frequency localization property and multi-resolution characteristics, which is in a wide range of applications in the field of signal processing. This paper studies the application of wavelet transform in signal filtering, by using an improved particle swarm optimization, proposes an intelligent signal de-noising method based on wavelet analysis. The method uses a Center Based Particle Swarm Algorithm (CBPSO) to select the optimal threshold for each sub-band in different scales, learning the type of noise from the signal itself intelligently, which does not require any prior knowledge of the noise. The improved particle swarm algorithm is used to enhance the optimal choice of the different scales of the wavelet domain threshold, which realized the signal De-noising under different types of noise background, and improved the speed of wavelet transform and wavelet construction, and has greater flexibility. The experimental results showed that CBPSO algorithm can get better De-noising effect