Can I convert one image in grayscale to monochrome?

Hello friends, I would know if there is chance to convert an grayscale image to monochrome

Answer 1

[Image Analyst]

A grayscale image is monochrome. There is nothing to do. There are no colors. Just look at the web site for any camera manufacturer out there, such as this really excellent camera from Lumenera: http://www.lumenera.com/products/industrial-cameras/lw11059.php. Gray scale images are monochrome, according to accepted and widely used industry terminology. (Of course a thresholded/binary image like Walter showed is also monochrome because it has no color, though I'd call them binary or logical since they don't have any grays in them other than pure black or pure white.)

[Image Analyst]

Here's some code to do the algorithm you asked for. I have doubts it's any good, and it's certainly inefficient, but it's what you asked for. If you want to share your image to get my advice on what to measure, feel free. I really have doubts that it's a good algorithm to do segmentation except for certain images. If you want to look at my File Exchange for some color segmentation methods, go here: http://www.mathworks.com/matlabcentral/fileexchange/?term=authorid%3A31862

clc;  % Clear command window.
clear;  % Delete all variables.
close all;  % Close all figure windows except those created by imtool.
workspace;  % Make sure the workspace panel is showing.
fontSize = 15;
% Read in a standard MATLAB color demo image.
folder = fullfile(matlabroot, '\toolbox\images\imdemos');
baseFileName = 'greens.jpg';
% Get the full filename, with path prepended.
fullFileName = fullfile(folder, baseFileName);
if ~exist(fullFileName, 'file')
  % Didn't find it there.  Check the search path for it.
  fullFileName = baseFileName; % No path this time.
  if ~exist(fullFileName, 'file')
    % Still didn't find it.  Alert user.
    errorMessage = sprintf('Error: %s does not exist.', fullFileName);
    uiwait(warndlg(errorMessage));
    return;
  end
end
rgbImage = imread(fullFileName);
% Get the dimensions of the image.  numberOfColorBands should be = 3.
[rows, columns, numberOfColorBands] = size(rgbImage);
% Display the original color image.
subplot(2, 2, 1);
imshow(rgbImage);
axis on;
title('Original Color Image', 'FontSize', fontSize);
% Enlarge figure to full screen.
set(gcf, 'units','normalized','outerposition',[0 0 1 1]);

% Convert to grayscale
grayImage = rgb2gray(rgbImage);
% Display the image.
subplot(2, 2, 2);
imshow(grayImage);
title('Gray Scale Image', 'FontSize', fontSize);

% Get an estimate of the background.
windowSize = 41;
backgroundImage = imfilter(double(grayImage), ones(windowSize)/windowSize^2);
% Display the image.
subplot(2, 2, 3);
imshow(backgroundImage, []);
title('Background Estimate', 'FontSize', fontSize);

% Subtract the background from the gray image.
backgroundSubtractedImage = double(grayImage) - backgroundImage;
% Display the image.
subplot(2, 2, 4);
imshow(backgroundSubtractedImage, []);
title('Background Subtracted Image', 'FontSize', fontSize);

[Manuel]

Image Analyst, I have no words,I just have to say thank you, very good, but there is a problem when you put imshow(backgroundSubtractedImage, []); the image appears normalized, but if you want to work with that image you can't, pic:"lena.tif"

[Image Analyst]

You're welcome, though like I said, I think that will be a poor algorithm in most situations, and would be willing to make some suggestions on an algorithm (if it's simple enough) if you want to post your image and say what you want to measure in the image and tell me how much the other images might differ from the one or two you posted.

[Walter Roberson]

"sorry we are unable to generate a view of the document at this time."

[Manuel]

https://docs.google.com/file/d/0BwumsnSY2aBoNVpsUEVoQWJEdEU/edit?usp=sharing

now you can see

[Image Analyst]

@Manuel that link doesn't work. Here is my code again with the input image attached and the output screenshot.

%============================================================================================================================================
% Initialization Steps.
clc;    % Clear the command window.
close all;  % Close all figures (except those of imtool.)
clear;  % Erase all existing variables. Or clearvars if you want.
workspace;  % Make sure the workspace panel is showing.
format long g;
format compact;
fontSize = 15;
markerSize = 4;

% Read in a standard MATLAB color demo image.
folder = fullfile(matlabroot, '\toolbox\images\imdemos');
baseFileName = 'greens.jpg';

% Get the full filename, with path prepended.
fullFileName = fullfile(folder, baseFileName);
if ~exist(fullFileName, 'file')
    % Didn't find it there.  Check the search path for it.
    fullFileName = baseFileName; % No path this time.
    if ~exist(fullFileName, 'file')
        % Still didn't find it.  Alert user.
        errorMessage = sprintf('Error: %s does not exist.', fullFileName);
        uiwait(warndlg(errorMessage));
        return;
    end
end
rgbImage = imread(fullFileName);

% Get the dimensions of the image.  numberOfColorBands should be = 3.
[rows, columns, numberOfColorBands] = size(rgbImage);
% Display the original color image.
subplot(2, 2, 1);
imshow(rgbImage);
axis on;
title('Original Color Image', 'FontSize', fontSize);
% Enlarge figure to full screen.
set(gcf, 'units','normalized','outerposition',[0 0 1 1]);
% Convert to grayscale
grayImage = rgb2gray(rgbImage);
% Display the image.
subplot(2, 2, 2);
imshow(grayImage);
title('Gray Scale Image', 'FontSize', fontSize);

% Get an estimate of the background.
windowSize = 41;
backgroundImage = imfilter(double(grayImage), ones(windowSize)/windowSize^2);
% Display the image.
subplot(2, 2, 3);
imshow(backgroundImage, []);
title('Background Estimate', 'FontSize', fontSize);

% Subtract the background from the gray image.
backgroundSubtractedImage = double(grayImage) - backgroundImage;
% Display the image.
subplot(2, 2, 4);
imshow(backgroundSubtractedImage, []);
title('Background Subtracted Image', 'FontSize', fontSize);

Like I said, it's no good but it's what you asked for originally (before you deleted your original question).

Answer 2

[Walter Roberson]

monochrome = (grayimage > some_threshold)

and see graythresh()

[Manuel]

sorry Walter but I dont understand, the grayscale image must have more gray levels than the image treshold? my idea is to convert an image to gray monochromatic, because this image monocrome want to subtract a filtered image average filter with a high intensity, to detect the centrelines, of this image

[Walter Roberson]

Are you wanting to compute the average intensity of a grayscale image? Or are you trying to convert a grayscale image to binary (pixels either on or off)?

[Manuel]

I will explain my problem, I have to do a segmentation process, for this situation I have to do monochromatic representation of the original color image who is normalised by subtract an estimate of background, witch is the result of a filtering operation with a large arithmetic mean kernel.

[Walter Roberson]

convert the original color image to double precision; convert the estimated background to double precision. Subtract to get the normalized version.

However, you cannot convert the normalized version to monochromatic by the usual calculation of luminance, as the calculation of luminance assumes that all the values are positive. You could calculate something, but it would not have any physical interpretation.

If you want to calculate the "something", then have a look at the documentation for rgb2gray() to find the formula that is used to convert RGB to luminance, and apply that formula to the normalized version that was calculated. You will get luminance-like values, but some of them will be negative and Science has yet to find a meaning for negative light. (How fast does negative light travel? -1*c ? Or should negative luminance be matched to tachyons ?)

Answer 3

[DGM]

One doesn't normalize data by subtraction. The process as described is common except for that one misunderstanding:

A = imread('printedtext.png');
A = im2double(A);
imshow(A)

bgest = imgaussfilt(A,30);
B = mat2gray(A./bgest); % use division

imshow(B)

This is essentially a local normalization process. To bring the image back into standard data range, I here chose to normalize the result using mat2gray().

An alternative would be to use imflatfield(). This would be essentially identical, except the result would be scaled by the mean of the background estimate instead of being normalized as with mat2gray(). The result is darker, but otherwise identical.

C = imflatfield(A,30);
imshow(C)

[Image Analyst]

I didn't know of the imflatfield() function. Thanks. I'll have to look into it.

You're right to use background division. Most people do background subtraction and that is not right except for a few cases like radiography and fluorescence.

As a final note I can see that my original answer up there has nothing at all to do with the (now) original post so it's very obvious that the original poster changed the post DRASTICALLY. I'll see if I can at least add some image and screenshot so that my answer makes sense by itself. I always attach the image and resulting screenshot so I'm not sure how they got removed. I often attach a text file with the author's post if it's the authors very first post since with first-time authors there is a very high probability they will obliterate their original post (like in this thread). The link @Manuel gave in a comment is also now broken.

Finally, I offer another algorithm to do document correction. It's meant to restore text, writing, or line diagrams on historical documents where the page has undergone color and brightness degradations.