基于LASSO算法的稀疏正则化高光谱图像的光谱解混合算法matlab仿真+仿真录像

clc; clear; close all; warning off; addpath(genpath(pwd)); size_x = 200; size_y = 200; size_lambda = 100; y = zeros(size_x,size_y,size_lambda); D = zeros(size_x*size_y, size_lambda); load('Data/celegans_spec_init.mat'); for i = 0:(size_lambda - 1) if i<10 temp = ['Stack000',num2str(i),'.tif']; else temp = ['Stack00',num2str(i),'.tif']; end img = imread(temp); temp_column = reshape(img,size_x*size_y,1); y(:,:,i+1) = img; D(:,i+1) = temp_column; end clear size_x size_y size_lambda temp temp_column i img; rmpath('Data/CElegans'); lambda = 0.6e1; copt = C_ML_Lasso( y, celegans_spec_init',lambda); sopt = celegans_spec_init; Comp1 = copt(:,:,1); Comp2 = copt(:,:,2); Comp3 = copt(:,:,3); Comp4 = copt(:,:,4); Comp1 = imadjust(uint8(255.*(Comp1 - min(Comp1(:) ))/(max(Comp1(:))-min(Comp1(:) )))); Comp2 = imadjust(uint8(255.*(Comp2 - min(Comp2(:) ))/(max(Comp2(:))-min(Comp2(:) )))); Comp3 = imadjust(uint8(255.*(Comp3 - min(Comp3(:) ))/(max(Comp3(:))-min(Comp3(:) )))); Comp4 = imadjust(uint8(255.*(Comp4 - min(Comp4(:) ))/(max(Comp4(:))-min(Comp4(:) )))); rgbComp1 = cat(3, Comp1*0.5, Comp1*0, Comp1*0.5); % magneta for oxidized lipid rgbComp2 = cat(3, Comp2*1, Comp2*0, Comp2*0); % red for nucleus rgbComp3 = cat(3, Comp3*0, Comp3*0.8, Comp3*0); % green for protein rgbComp4 = cat(3, Comp4*0, Comp4*0.3, Comp4*0.3);% cyan for water C_disp = rgbComp1+rgbComp2+rgbComp3+rgbComp4; figure; imshow(C_disp,'InitialMagnification',150); Spec1 = sopt(1,:); Spec2 = sopt(2,:); Spec3 = sopt(3,:); Spec4 = sopt(4,:); figure; plot(Spec1,'m','Linewidth',1.5); hold on; plot(Spec2,'r','Linewidth',1.5); plot(Spec3,'g','Linewidth',1.5); plot(Spec4,'c','Linewidth',1.5); xlabel 'Channel'; ylabel 'Int. (a.u.)'; legend('Oxidized lipid','Fat droplets','lysosome-related organelles','Protein') pbaspect([2 1 1]);