基于GA和MP的信號稀疏分解matlab程序.doc

基于GA和MP的信号稀疏分解 %% MATLAB script for MP algorithm based on FFT % % %******************** Preparation part ***************************** % %************************** part 1:imput signal *************************** Im=wavread(FEM); Im=Im; %**************************** part 2: set parameters *********************** f1=50; f2=160; fs=200; iterative_number=100; % the matching pursuit processing iterative number [a,N]=size(Im); % the longth of signal and the longth of atoms; N a_base=2; % scale: the scale is determined by a and j; j_min=0; j_max=log2(N); u_base=1/2; % the transmission or displacement u should determined as following v_base=pi; % the frequency v k_min=0; w_base=pi/6; % the phase w i_min=0; i_max=12; %%************************ initialization ******************************** signal=Im; signal_reconstruct=zeros(1,N); signal_r=Im; %*************** wipe off the direct current vector ************************* signal_reconstruct=(1/N)*sum(signal); signal_r=signal-signal_reconstruct; %*********************start calculation********************************* for n=1:iterative_number % ********the following program uses one subroutine to select the best atom************** [proj,scale,translation,freq,phase]=gas(signal_r,N,a_base,j_min,j_max,u_base,p_min,v_base,k_min,w_base,i_min,i_max); %******reconstruct the best atom from the parameters got by above subroutine********* t=0:N-1; t=(t-translation)/scale; g=(1/sqrt(scale))*exp(-pi*t.*t).*cos(freq*t+phase); g=g./sqrt(sum(g.*g)); %normalizatiom the atom %***********reconstruct signal and renew the residual********************** signal_reconstruct=signal_reconstruct+proj*g; signal_r=signal_r-proj*g; %************************* the figure of result ************************ subplot(221); plot(signal); % original siganl subplot(222); plot(g); %the atom subplot(223); plot(signal_r); %the residual subplot(224); plot(signal_reconstruct); %reconstruct signal end % *************