Landscape paper coding (v1_8august2016)

clear;
clc;
nlist1 = {'all_used/1.png'};
nlist2 = {'all_used/2.png'};
nlist3 = {'all_used/3.png'};
nlist4 = {'all_used/4.png'};
nlist5 = {'all_used/5.png'};
nlist6 = {'all_used/6.png'};
nlist7 = {'all_used/7.png'};
nlist8 = {'all_used/8.png'};
nlist9 = {'all_used/9.png'};
nlist10 = {'all_used/10.png'};

shape_value = 2.25;
isolation_value = 0.4;
isolation_value_based = 0.2;
picture = nlist1; % Choose for test picture.

% test function kmeans
X = [randn(20,2)+ones(20,2); randn(20,2)-ones(20,2)];
      opts = statset('Display','final');
      [cidx, ctrs] = kmeans(X, 4, 'Distance','city', ...
                            'Replicates',5, 'Options',opts);

for cta = 1:1 % all in my img folder is 7
    nm = picture{cta};
    rgb = imread(nm);
    normImage = im2double(rgb);
    nm_input = rgb2gray(normImage);

% Step 1: Read Image
    img_resize = imresize(rgb, 1);
    he = img_resize;
    he2 = img_resize;
 %  figure, imshow(he), title('1'); % Showing a original image.

    img_graythresh = rgb2gray(he );
    threshold = graythresh(img_graythresh);
    bw_img = im2bw(he,threshold);
%   figure, imshow(bw_img), title('2');
%  figure, imhist(img_graythresh), title('3');

% Step 2: Convert Image from RGB Color Space to L*a*b* Color Space
    cform = makecform('srgb2lab');
    cform_rgb = makecform('lab2srgb');
    lab_he = applycform(he,cform);
    rgb_he = applycform(he,cform_rgb);

% Step 3: Classify the Colors in 'a*b*' Space Using K-Means Clustering
    ab = double(rgb_he(:,:,2:3));
    ab_all_band = double(rgb_he(:,:,2:3));
    nrows = size(ab,1);
    ncols = size(ab,2);
    ab = reshape(ab,nrows*ncols,2);

    nColors = 3;
    [cluster_idx, cluster_center] = kmeans(ab,nColors, ...
                                    'Replicates',1, ...
                                    'Options',opts, ...
                                    'start',[118 143; 127 140; 133 127]);
                                % 'start',[118 143; 127 140; 133 127]);

% Step 4: Label Every Pixel in the Image Using the Results from KMEANS
    pixel_labels = reshape(cluster_idx,nrows,ncols);
    %figure(2), imshow(pixel_labels,[]), title('image labeled by cluster index')

% Step 5: Create Images that Segment the H&E Image by Color.
    segmented_images = cell(1,3);
    rgb_label = repmat(pixel_labels,[1 1 3]);

    for k = 1:3
         color = rgb;
         color(rgb_label ~= k) = 0;
         segmented_images{k} = color;
    end

    %figure, imshow(segmented_images{1}), title('objects in cluster 1')
    %figure, imshow(segmented_images{2}), title('objects in cluster 2')
    figure, imshow(segmented_images{3}), title('objects in cluster 3')  % Show only grey cluster (road cluster)

    %Normaliztion for clustering road area.
    normImage_c3 = mat2gray(segmented_images{3});
    normImage_c3_gray =rgb2gray(segmented_images{3});
    %figure, imhist(normImage_c3_gray);

    % to Binary image
    gray_convert_image = rgb2gray(normImage_c3);
    binary_convert_image = im2bw(gray_convert_image);
    figure, imshow(binary_convert_image)
    bwarea(binary_convert_image);

    %bw = imread('text.tif');
    %se = strel('line',20,150);
    se = strel('line',10,10);
    bw2 = imdilate(binary_convert_image,se);% graythresh = 'bw_img' % kmeans = 'binary_convert_image'
    %imshow(bw), title('Original')
    figure, imshow(bw2), title('Dilated')
    bw3 = ~bw2;
    bw4 = im2bw(bw3);
    %figure, imshow(bw4);
    CC = bwconncomp(bw2); %Connected Component Finding all area.
    %CC = bwconncomp(my_image);
    L = labelmatrix(CC);
    A = cell( size(CC.PixelIdxList,1) , size(CC.PixelIdxList,2) );
    A = CC.PixelIdxList;
    size(CC.PixelIdxList,2);

    %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%


    %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

    La=bwlabel(bw2,8); %%% labeledImage is a binary image
    figure,imshow(La,[]);
    coloredLabel = label2rgb(La, 'hsv', 'k', 'shuffle');
    imshow(coloredLabel);

  %  pr = regionprops( La, 'Area', 'PixelIdxList' );

    %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

    stats = regionprops(bw2,'Area')
    stats2 = regionprops(bw2,'Centroid')


    %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
    smallAreaA = La;
    smallAreaSh = La;
    smallArea = La;
    smallArea2 = La;
    smallArea3 = La;
    smallArea4 = La;
    smallArea5 = La;
    s = regionprops(La, 'Orientation','Area','PixelIdxList','Perimeter', 'BoundingBox',...
        'MajorAxisLength', 'MinorAxisLength', 'Eccentricity', 'Centroid');

    %%%%%%%%%%%%%%%%%%%%%%%  Regionprop to matrix %%%%%%%%%%%%%%%%%%%%%%%%%
    num_boject = 1:1:size(s);
    s_centroid = vertcat(s.Centroid);
    cen1 = s_centroid(:,1);
    cen2 = s_centroid(:,2);
    region_matrix = [s.Area; s.Perimeter; num_boject;]';

    s_area = region_matrix(:,1);
    s_perimeter = region_matrix(:,2);
    s_num = region_matrix(:,3);
    min_perimeter = (sqrt(region_matrix(:,1)))*4;
    shape_index = s_perimeter./min_perimeter; % formula of shape index
    % region_matrix_used will contain 1. number of objects 2. perimeter and
    % min_perimeter and calculating to shape_index
    region_matrix_used = [s_num s_area s_perimeter min_perimeter shape_index];
    threshold_use = max(s_area)/2;

    score = shape_index;
    score2 = s_area;
    num2 = 0;
    area2 = 0;
    shape2 = 0;
    cent1 = 0;
    cent2 = 0;
    j = 1;
    k = 1;

    % Loop 1 for "region_state1"
    for i=1:size(s)
        if s_area(i) > threshold_use & shape_index(i) > shape_value;
           q1 = s_num(i);
           q2 = s_area(i);
           q3 = shape_index(i);
           q4 = cen1(i);
           q5 = cen2(i);
           num(j) = q1;
           area(j) = q2;
           shape(j) = q3;
           centA(j) = q4;
           centB(j) = q5;
           j = j+1;
        end
    end

     % Loop 2 for "region_state2"
    for i=1:size(s)
        if s_area(i) < threshold_use & shape_index(i) > shape_value;
           q1 = s_num(i);
           q2 = s_area(i);
           q3 = shape_index(i);
           q4 = cen1(i);
           q5 = cen2(i);
           num2(k) = q1;
           area2(k) = q2;
           shape2(k) = q3;
           cent1(k) = q4;
           cent2(k) = q5;
           k = k+1;
        end
    end
    region_state1 = [num' area' shape' centA' centB']; % Find Main Road
    region_state2 = [num2' area2' shape2' cent1' cent2']; % Find Local Road

    temp1 = 1;
    all_object = max(size(region_state2(region_state2(:,3)>0)));
         area = region_state1(:,2);
         cen_road1 = region_state1(:,4);
         cen_road2 = region_state1(:,5);
         cen_check1 = region_state2(:,4);
         cen_check2 = region_state2(:,5);

  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  %%%%%%%%%%%%%%%%%%%%%%% Formula of Isolation Index %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
    isolation_form  = 0;
    num_main_road = size(num',1)
    num_local_road = size(num2',1)
       for local=1:num_local_road   %% i = number of checking for local road.
           for main=1:num_main_road
                 distance = [cen_road1(main),cen_road2(main);cen_check1(local),cen_check2(local)];
                 dis = pdist(distance,'euclidean');
                 isolation_form = isolation_form +area(main)/(dis^2);
           end
           similarity(local) = isolation_form / num_main_road ;
           isolation_form = 0;
       end
  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

    region_state3 = [num2' area2' shape2' similarity']; % Calculating to be a similarity index

    iso_num = region_state3(:,1);
    iso_iso = region_state3(:,4);
    indexIso=1;

    % This loop was created for keep similarity index to matrix.
    for i=1:size(s_num)
       Test(i) = 0;
       for j=1:num_local_road  % take isolation following numners of it.
           if i == iso_num(j)
               Test(i) = iso_iso(j);
           end
       end


    end

    region_state4 = [s_num Test']; % Created for checking above loop.
    isolate = region_state4(:,2);


    threshold_max = max(region_matrix_used(:,2))/2

    small_select = ((s_area >= threshold_max) & (shape_index > shape_value));
    smallArea( vertcat( s(~small_select).PixelIdxList ) ) = 0; %// set all other regions to zero

    % This is Main Road.
    MainRoad = (s_area >= threshold_max) & (shape_index > shape_value);

    small_select3 = ((shape_index > shape_value)); %all
    smallArea3( vertcat( s(~small_select3).PixelIdxList ) ) = 0;

    small_select4 = (isolate > isolation_value_based);
    smallArea4( vertcat( s(~small_select4).PixelIdxList ) ) = 0;

    % This is Local Road.
    LocalRoad = (isolate > isolation_value);

    % This is Results.
    small_select5 = (MainRoad + LocalRoad);
    smallArea5( vertcat( s(~small_select5).PixelIdxList ) ) = 0;


    figure, imshow( smallArea ); colormap( summer ); % Main_Road Result
    figure, imshow( smallArea3 ); %all
    figure, imshow( smallArea4 ); colormap( summer ); % Local_Road Result
    figure, imshow( smallArea5 ); % Correct_Road Result

    figure, imshow(nm), hold on, himage = imshow(smallArea5), set(himage, 'AlphaData', 0.6);


    %%%%%%%%%%%%%%%%%%%%%% Thining the image %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
    % Thining the image
    skeletionization_image = bwmorph(smallArea5,'thin',Inf);
    se2 = strel('line',2,2); % 'line',5,8  'best',2,2
    bw2 = imdilate(skeletionization_image,se2);
    coloredLabel1 = label2rgb(bw2, 'hsv', 'k', 'shuffle');
    % figure, imshow(coloredLabel1);
    % ps     = dpsimplify(skeletionization_image,1);

    % SHOW ---------------- Thining the image ---------------------

    figure, imshow( smallArea5 ); colormap( summer ), hold on
    himage = imshow(coloredLabel1);
    set(himage, 'AlphaData', 0.7);

%     %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%
%
%     %%%%%%%%%%%%%%% Skeleton the image (Not using in this paper) %%%%%%%
%
%     % Skeleton the image
%     skeletionization_image = bwmorph(smallArea,'skel',Inf);
%     se2 = strel('line',2,2); % 'line',5,8  'best',2,2
%     bw2 = imdilate(skeletionization_image,se2);
%     coloredLabel2 = label2rgb(bw2, 'hsv', 'k', 'shuffle');
%     % figure, imshow(coloredLabel1);
%     % ps     = dpsimplify(skeletionization_image,1);
%
%     % SHOW ---------------- Skeleton the image ---------------------
%     figure, imshow(nm), hold on
%     himage = imshow(coloredLabel2);
%     set(himage, 'AlphaData', 0.7);
%
    %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%


    %%%%%%%%%%%%%%%%%%%%%% Simplipfy with DP %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

%  - - - - - - - - - - - - - Find brach I - - - - - - - - - - - - - - - -
    skelImg   = bwmorph(smallArea5, 'thin', 'inf');
    branchImg = bwmorph(skelImg, 'branchpoints');
    endImg    = bwmorph(skelImg, 'endpoints');

    [row, column] = find(endImg);
    endPts        = [row column];
    [row, column] = find(branchImg);
    branchPts     = [row column];

    figure; imshow(skelImg); hold on;
    plot(branchPts(:,2),branchPts(:,1),'r*'); hold on;
    plot(endPts(:,2),endPts(:,1),'*');

  % - - - - Simplipfy with DP - - Find branch II - deleting some node - - %
    skel2= bwmorph(smallArea5,'skel',Inf);

    B = bwmorph(skel2, 'branchpoints');
    E = bwmorph(skel2, 'endpoints');

    [y,x] = find(E);
    B_loc = find(B);

    Dmask = false(size(skel2));
    for k = 1:numel(x)
        D = bwdistgeodesic(skel2,x(k),y(k));
        distanceToBranchPt = min(D(B_loc));
        Dmask(D < distanceToBranchPt) =true;
    end
    skelD = skel2 - Dmask;
    coloredLabel2 = label2rgb(skelD, 'hsv', 'k', 'shuffle');
  %  skelC     = dpsimplify(skelD, 0.5); %Simplipfy with DP
    figure, imshow(skelD);
    hold all;
    [y,x] = find(B); plot(x,y,'ro')

   % figure, imshow(nm), hold on
    figure, imshow( smallArea5 ); colormap( summer ), hold on % my output
    himage = imshow(coloredLabel2);
    set(himage, 'AlphaData', 0.7);


    %     figure, imshow( smallArea ), hold on % my output
    %     himage = imshow(coloredLabel2);
    %     set(himage, 'AlphaData', 0.7);

    %  - - - - - - - -  Simplipfy with DP (Uisng function dpsimplify) - - - - - - - - - -
    %
%     B2 = num2cell(B);
%     tol    = 0.5;
%     ps     = dpsimplify(B2,tol);
%     hold on
%     figure, plot(ps(:,1),ps(:,2),'r','LineWidth',2);
%     legend('original polyline','simplified')


    %%%%%%%% Pseudocode (refer to
    %%%%%%%% https://en.wikipedia.org/wiki/Ramer%E2%80%93Douglas%E2%80%93Peucker_algorithm)
    %%%%%%%%
    % function DouglasPeucker(PointList[], epsilon)
    %     // Find the point with the maximum distance
    %     dmax = 0
    %     index = 0
    %     end = length(PointList)
    %     for i = 2 to ( end - 1) {
    %         d = perpendicularDistance(PointList[i], Line(PointList[1], PointList[end]))
    %         if ( d > dmax ) {
    %             index = i
    %             dmax = d
    %         }
    %     }
    %     // If max distance is greater than epsilon, recursively simplify
    %     if ( dmax > epsilon ) {
    %         // Recursive call
    %         recResults1[] = DouglasPeucker(PointList[1...index], epsilon)
    %         recResults2[] = DouglasPeucker(PointList[index...end], epsilon)
    %
    %         // Build the result list
    %         ResultList[] = {recResults1[1...length(recResults1)-1], recResults2[1...length(recResults2)]}
    %     } else {
    %         ResultList[] = {PointList[1], PointList[end]}
    %     }
    %     // Return the result
    %     return ResultList[]
    % end


 end