Wiener 滤波器的信号去噪算法研究附matlab代码.zip

function [xest,B,MSE] = wienerFilt(x,y,N) % % Wiener filter based on Wiener-Hopf equations % This function takes as inputs a noisy signal, x, and a reference signal, y, % in order to compute a N-order linear filter that provides an estimate of y % from x % % INPUTS % x = noisy signal % y = reference signalsinit % N = filter order % % OUTPUTS % xest = estimated signal % b = Wiener filter coefficents % MSE = mean squared error % X = 1/N .* fft(x(1:N)); Y = 1/N .* fft(y(1:N)); X = X(:); Y = Y(:); Rxx = N .* real(ifft(X .* conj(X))); % Autocorrelation function Rxy = N .* real(ifft(X .* conj(Y))); % Crosscorrelation function Rxx = toeplitz(Rxx); Rxy = Rxy'; B = Rxy / Rxx; B = B(:); % Wiener-Hopf eq. B = inv(Rxx) Rxy xest = fftfilt(B,x); xest = xest(N+1:end); % cut first N samples due to distorsion during filtering operation MSE = mean(y(N+1:end) - xest) .^2; % mean squared error