detect horizontal and vertical lines

How to detect horizontal and vertical lines in a given image using imerode function? I have this code but it doesn't detect H and V lines

im=imread('src\ima.tif');
im(:,:,2:4)=[];
im=im2bw(im);
SE = strel('arbitrary',ones(10,1));
im2 = imerode(im,SE);
imwrite(im2,'src\aa.tif');
imshow(im2)

Answer 1

[Image Analyst]

Just call bwconncomp(), then regionprops() and check if the bounding box height is less than some number, like 1 or 2 pixels. Untested code:

binaryImage = grayImage < 128;
cc = bwconncomp(binaryImage);
measurements = regionprops(cc, 'BoundingBox');
for k = 1 : length(measurements)
  thisBB = measurements(k).BoundingBox;
  if thisBB(4) <= 3  % If it's shorter than 4 lines tall.
    message = sprintf('Blob #%d is horizontal.', k);
    sprintf('%s\n', message);
    uiwait(helpdlg(message));
  end
end

[Mohammad]

thanks but this doesn't do what I want. As I said i want to detect the H & V lines and create a new image contains these V & H lines.

[Image Analyst]

Not, not completely. I was hoping this little bit was enough for you to finish it. I did not do a complete turnkey application for you. All you have to do to do what you want is to put imcrop into the loop

croppedImages{k} = imcrop(binaryImage, thisBB);

and to change the if statement to look for thisBB(3) less than some width

% If it's shorter than 4 lines tall or narrower than 3 lines.
aspectRatio = thisBB(4)/thisBB(3);
if (thisBB(4) <= 3 || thisBB(3) <= 3) && aspectRatio > 4

I check the aspect ratio to make sure that it must be long and skinny, not just a little round dot or something. If you still can't complete the program, let me know. You might have to debug or change the numbers somewhat. I did not test this - it's just off the top of my head, based on my experience.

[Mohammad]

Thanks again. I am new to Matlab and I am trying to learn from your code which will be like the following:

clc
grayImage =imread('src\a.tif');
binaryImage = grayImage < 128;
cc = bwconncomp(binaryImage);
measurements = regionprops(cc, 'BoundingBox');
for k = 1 : length(measurements)
  thisBB = measurements(k).BoundingBox;
  aspectRatio = thisBB(4)/thisBB(3);
  if (thisBB(4) <= 3 || thisBB(3) <= 3) && aspectRatio > 4
    croppedImages{k} = imcrop(binaryImage, thisBB);
  end
end

But again how to show this image "croppedImages" I tried imshow and imwrite but it doesn't work?

[Image Analyst]

Alright. You just forgot to call imshow after you called imcrop. But I did it for you. Below is a full blown demo:

clc;    % Clear the command window.
close all;  % Close all figures (except those of imtool.)
imtool close all;  % Close all imtool figures if you have the Image Processing Toolbox.
clear;  % Erase all existing variables. Or clearvars if you want.
workspace;  % Make sure the workspace panel is showing.
format short g;
format compact;
fontSize = 20;
folder = 'C:\Users\Mohammad\Documents\Images';
baseFileName = 'ima.jpg';
% Get the full filename, with path prepended.
fullFileName = fullfile(folder, baseFileName);
% Check if file exists.
if ~exist(fullFileName, 'file')
  % File doesn't exist -- didn't find it there.  Check the search path for it.
  fullFileNameOnSearchPath = baseFileName; % No path this time.
  if ~exist(fullFileNameOnSearchPath, 'file')
    % Still didn't find it.  Alert user.
    errorMessage = sprintf('Error: %s does not exist in the search path folders.', fullFileName);
    uiwait(warndlg(errorMessage));
    return;
  end
end
grayImage = imread(fullFileName);
% Save this figure handle.
hFig1 = gcf;
% Get the dimensions of the image.  
% numberOfColorBands should be = 1.
[rows, columns, numberOfColorBands] = size(grayImage);
if numberOfColorBands > 1
  % It's not really gray scale like we expected - it's color.
  % Convert it to gray scale by taking only the green channel.
  grayImage = grayImage(:, :, 2); % Take green channel.
end
% Display the original gray scale image.
subplot(1, 3, 1);
imshow(grayImage, []);
title('Original Grayscale Image', 'FontSize', fontSize);
% Enlarge figure to full screen.
set(gcf, 'Units', 'Normalized', 'OuterPosition', [0 0 1 1]);
% Give a name to the title bar.
set(gcf, 'Name', 'Demo by ImageAnalyst', 'NumberTitle', 'Off') 

binaryImage = grayImage < 128;
% Display the binary image.
subplot(1, 3, 2);
imshow(binaryImage, []);
title('Binary Image', 'FontSize', fontSize);
% Display the binary image again where we can put boxes over it.
subplot(1, 3, 3);
imshow(binaryImage, []);
hold on;
title('Binary Image', 'FontSize', fontSize);

% Create a figure for the cropped images.
hCropped = figure;

% Do connected components analysis on it.
cc = bwconncomp(binaryImage);
% Measure the bounding box of all blobs.
measurements = regionprops(cc, 'BoundingBox');
fprintf('Found %d regions\n', cc.NumObjects);
numSkinnyRegions = 0;
for k = 1 : cc.NumObjects
  figure(hFig1); % Switch to figure 1.
  thisBB = measurements(k).BoundingBox
  % Draw a box around the region in cyan.
  hRect = rectangle('Position', thisBB, 'EdgeColor', 'c', 'LineWidth', 3);
  aspectRatio(k) = thisBB(4)/thisBB(3);
  if (thisBB(4) <= 3 || thisBB(3) <= 3) && (aspectRatio(k) > 4 || aspectRatio(k) < 1/4)
    numSkinnyRegions = numSkinnyRegions + 1;
    % Save it to a cell array, just in case we want to use it after the loop is done.
    croppedImages{numSkinnyRegions} = imcrop(binaryImage, thisBB);    
    % Draw skinny regions in a different color
    delete(hRect); % Get rid of old one.
    hRect = rectangle('Position', thisBB, 'EdgeColor', 'r', 'LineWidth', 3);  
    % Switch to figure 2
    figure(hCropped);
    subplot(2, 3, numSkinnyRegions);
    imshow(croppedImages{numSkinnyRegions}, []);
    caption = sprintf('Blob #%d', numSkinnyRegions);
    title(caption, 'FontSize', fontSize);    
  end
end
% Enlarge figure to full screen.
set(hCropped, 'Units', 'Normalized', 'OuterPosition', [0 0 1 1]);
% Give a name to the title bar.
set(hCropped, 'Name', 'Demo by ImageAnalyst', 'NumberTitle', 'Off')

[Mohammad]

Thanks very much

[Image Analyst]

You're welcome. Since it seems like it does everything you want, would you mind going ahead and marking my answer as Accepted? Thanks in advance.

[Tina]

Hi, I'm new to Matlab and Matlab Answers. I have a query. While detecting vertical lines, what if I need to detect the lines from the alphabets also? For example quoting the same input image as above, the letter H has two vertical lines, letters R, L, D has one vertical line. How do I detect them? Thanks.

[Firdausy angga Permana]

Hi Image Analyst. I've tried this code and it's really helping me to solve my problem. But, how can we remove that lines? Thanks Before!