Comparing multiple images ( via edge detection or strongest gradient between dark and light)

Hello guys,

first of all : i am very new to matlab so dont blame me for asking.

I want to compare some pictures ( 5 pictures or so ).I am testing a light,every hour i take a photo and then compare them to see if the'cut-off' did move. And if the 'cut off' moves, I'd like to see how much it shifted.

Every little adive would help me !

( I uploadet the picutre. In my test, the pictures will loke the same. But maybe afters 3 hours the 'cut-off will shift minimal. And thats what i wanna se. Like if i could get a number in the end like 2 == It shifted 2 millimeters.)

THX GUYS !

[Rik]

Question recovered from Google cache:

Comparing multiple images ( via edge detection or strongest gradient between dark and light)

Hello guys,

first of all : i am very new to matlab so dont blame me for asking.

I want to compare some pictures ( 5 pictures or so ).I am testing a light,every hour i take a photo and then compare them to see if the'cut-off' did move. And if the 'cut off' moves, I'd like to see how much it shifted.

Every little adive would help me !

( I uploadet the picutre. In my test, the pictures will loke the same. But maybe afters 3 hours the 'cut-off will shift minimal. And thats what i wanna se. Like if i could get a number in the end like 2 == It shifted 2 millimeters.)

THX GUYS !

[Rena Berman]

(Answers Dev) Restored edit

Answer 1

[Image Analyst]

I don't know what the "cut-off" is. Is it the shadow line? Do you want the row number of the shadow for every column in the image? Something like

grayImage = imread(filename);
[rows, columns, numberOfColorChannels] = size(grayImage)
if numberOfColorChannels == 3
    grayImage = rgb2gray(grayImage);
end
binaryImage = grayImage < someValue; % For example someValue = 100 or whatever.
shadowRows = zeros(1, columns);
for col = 1 : columns
    t = find(binaryImage(:, col), 1, 'first')
    if ~isempty(t)
        shadowRows(col) = t;        
    end
end
plot(shadowRows, 'b-')
grid on;

Answer 2

[Image Analyst]

Try this:

clc;    % Clear the command window.
fprintf('Beginning to run %s.m.\n', mfilename);
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 = 20;

baseFileName1 = 'hdg.png';
fullFileName1 = fullfile(pwd, baseFileName1);
grayImage1 = imread(fullFileName1);
[rows, columns, numberOfColorChannels] = size(grayImage1)
if numberOfColorChannels == 3
    grayImage1 = rgb2gray(grayImage1);
end
% Display image.
subplot(2, 3, 1);
imshow(grayImage1, []);
impixelinfo
title(baseFileName1, 'FontSize', fontSize);

% Threshold the image.
thresholdValue = 125;
binaryImage = grayImage1 < thresholdValue;
% Take largest blob.
binaryImage = bwareafilt(binaryImage, 1); 
% Do a morphological opening to get rid of grid lines.
binaryImage = imopen(binaryImage, true(3)); 
% Display image.
subplot(2, 3, 4);
imshow(binaryImage, []);
impixelinfo
title(baseFileName1, 'FontSize', fontSize);

shadowRows1 = zeros(1, columns);
for col = 1 : columns
    t = find(binaryImage(:, col), 1, 'last')
    if ~isempty(t)
        shadowRows1(col) = t;        
    end
end

baseFileName2 = 'hdg1.png';
fullFileName2 = fullfile(pwd, baseFileName2);
grayImage2 = imread(fullFileName2);
[rows, columns, numberOfColorChannels] = size(grayImage2)
if numberOfColorChannels == 3
    grayImage2 = rgb2gray(grayImage2);
end
% Display image.
subplot(2, 3, 2);
imshow(grayImage2, []);
impixelinfo
title(baseFileName2, 'FontSize', fontSize);

% Threshold the image.
binaryImage = grayImage2 < thresholdValue;
% Take largest blob.
binaryImage = bwareafilt(binaryImage, 1); 
% Do a morphological opening to get rid of grid lines.
binaryImage = imopen(binaryImage, true(3)); 
% Display image.
subplot(2, 3, 5);
imshow(binaryImage, []);
impixelinfo
title(baseFileName2, 'FontSize', fontSize);

shadowRows2 = zeros(1, columns);
for col = 1 : columns
    t = find(binaryImage(:, col), 1, 'last');
    if ~isempty(t)
        shadowRows2(col) = t;        
    end
end

subplot(2, 3, 3);
plot(shadowRows1, 'b-')
hold on;
plot(shadowRows2, 'r-')
grid on;
xlabel('Row', 'FontSize', fontSize);
ylabel('Column', 'FontSize', fontSize);
title('Shadow Location', 'FontSize', fontSize);
legend('image 1', 'image 2', 'Location', 'northwest');

% The images are not the same size for some strange reason so we can't subtract the arrays until we match their sizes
if length(shadowRows1) > length(shadowRows2)
	shadowRows2(length(shadowRows1)) = shadowRows2(end); % Extend last value out.
else
	shadowRows1(length(shadowRows2)) = shadowRows1(end); % Extend last value out.
end
% Plot the difference.
diffRows = shadowRows1 - shadowRows2;
subplot(2, 3, 6);
plot(diffRows, 'b-')
grid on;
xlabel('Row', 'FontSize', fontSize);
ylabel('Vertical Difference', 'FontSize', fontSize);
title('Difference in Shadow Locations', 'FontSize', fontSize);

% Maximize figure.
g = gcf;
g.WindowState = 'maximized'

[Image Analyst]

I didn't know that you didn't have any control over the exposure. Instead of manual thresholding:

% Threshold the image.
binaryImage = grayImage2 < thresholdValue;

Try automatic thresholding:

% Threshold the image.
binaryImage = imbinarize(grayImage2);

[Image Analyst]

Rows go down, not up. Row 1 is at the top, not the bottom like in your plot. To get them to go in the same directions after you call plot(), call axis ij

plot(x, y);
axis ij;

The imbinarize() uses the Otsu algorithm and its algorithm can be found online. If you want to adjust the location (threshold), then I think there are some input options to imbinarize() that you can use to move the line up or down.

[Image Analyst]

You put the axis ij command after the plot() command:

% Plot the difference.
diffRows = shadowRows1 - shadowRows2;
subplot(2, 3, 6);
plot(diffRows, 'b-')
axis ij;
grid on;
xlabel('Row', 'FontSize', fontSize);
ylabel('Vertical Difference', 'FontSize', fontSize);
title('Difference in Shadow Locations', 'FontSize', fontSize);

[Arb2493]

thx for the response.

I have a problem with these pictures. The code doesnt recognize the shadow line ?

[madhan ravi]

You must be kidding right? You deleted the question!!