Find Image/Reference Image within Image

Hi All, I need to find the icons displayed in a pdf file and output them in a tabular format. I have turned the pdf file to png to be able to do some image processing on the image.

I need to use the reference icons to detect the status of the vehicle for each test and output that result in table format. The tests vary depending if it is a mechanical report or it states its a hybrid report therefore there may sometimes be additional columns to analyse. I have also attached the png versions of the reference icons on their own.

I require a table such as this example;

(col1) "Green" ,(col2) "Green" , (col3) "Orange" etc.....

It would be additionally helpul if it could run through each of the files also and output the results in a master table if possible.

[yanqi liu]

yes,sir，may be use color and shape to make object match

Answer 1

[Jaynik]

Hi Taru,

I assume that the specification sheet template remains the same. If that is the case, instead of using image processing or machine learning algorithms, you can simply set a template using coordinates of each icon.

As you have different images for the hybrid and mechanical report, you can easily determine whether to generate a column for hybrid or mechanical. Below is a sample code you can use to get started. I assume the "Standard Mechanical Report" for now:

finalAns = strings(2,3);
col1_top = [426, 243];  % Setting coordinates of each icon
col1_bottom = [434, 800];   % These should be set appropriately
col2_top = [891, 262];
col2_bottom = [895, 552];
col3_top = [1119, 259];
col3_bottom = [1119, 542];

coordsToCheck = [col1_top; col1_bottom; col2_top; col2_bottom; col3_top; col3_bottom];
img = imread("sheet1.png"); % Name of your specs seet

% if else statement to check if mechanical or hybrid

for i=1:size(coordsToCheck, 1)
    y = ceil(i / size(finalAns, 1));
    x = i - (y - 1) * size(finalAns, 1);
    rgbValues = double(squeeze(img(coordsToCheck(i,2), coordsToCheck(i,1), :))) / 255;
    rgbArray = reshape(rgbValues, [1, 1, 3]);
    hsv = rgb2hsv(rgbArray);
    hueDegrees = hsv(1) * 360;
    saturationPercent = hsv(2) * 100;
    valuePercent = hsv(3) * 100;
    % Set colors based on hue and rgb values
    if (hueDegrees <= 30 && hueDegrees >= 25)
        finalAns(x,y) = "Orange";
    elseif (hueDegrees <= 92 && hueDegrees >= 75)
            finalAns(x,y) = "Green";
    elseif (hueDegrees >= 345)
        finalAns(x,y) = "Red";
    elseif (rgbValues(1) == 0)
        finalAns(x,y) = "Black";
    else
        finalAns(x,y) = "White";
    end
end

I hope this works for you, feel free to reach out if you have any other queries regarding the code.