【图像重建】基于OMP算法实现二维图像压缩感知图像重建（含PSNR）附Matlab代码.zip

function out = CS(X, const_C, BLOCK_SIZE, INDEX, sparse_level, NUM) [M, N] = size(X); out = zeros(M,N); B = BLOCK_SIZE; block_size = B * B; m = M /B; n = N / B; block_num = m * n; Z = ones(M, N); mesur_num = zeros(1, m*n); ind = 1; for i = 1: m for j = 1: n temp1 = zeros(B, B); temp1 = X(B*(i-1)+1 : B*i, B*(j-1)+1 : B*j); Y = dct2(temp1); %计算该块中0系数的个数 % Y1 = abs(Y); % Y2 = Y1 * 255; % Y3 = round(Y2); % Y3 = round(Y1); % temp_matrix = find(abs(Y3) < 1); % [count1, count2] = size(temp_matrix); % count = count1 * count2; % matrix1D(ind) = count; %非0个数 即稀疏度 %MER_NUM = round(const_C*sparse_level*log(block_size/sparse_level)); %有待调整 MER_NUM = NUM; mesur_num(ind) = MER_NUM; ind = ind + 1; %merix to vector vector_X = matrix2vector(Y, B, B, INDEX); X1 = vector_X'; %CS mer_matrix = randn(MER_NUM, B*B); Y4 = mer_matrix * X1; %OMP RE_VECTOR = CS_OMP(block_size, Y4, mer_matrix, sparse_level); %vector to matrix RE_MATRIX = vector2matrix(RE_VECTOR, B, B, INDEX); S = idct2(RE_MATRIX); Z(B*(i-1)+1 : B*i, B*(j-1)+1 : B*j) = S; end; end; T_M = 0; for i =1:block_num T_M = T_M + mesur_num(i); end; M_R = T_M/(M*N) out = Z;