How do I analyze color uniformity of my image in Matlab?
19 Ansichten (letzte 30 Tage)
Ältere Kommentare anzeigen
I have a image of a paper sensor made by dipping in a bunch of solutions. After drying, it produces almost uniform color all over the paper. But is there any way I can analyze this image to arrive at a perticular value of uniformity? My ultimate goal is to be able to produce uniform color and be able to reproduce same results that I can put in matlab and confirm the value. I have attached the image that I want to analyze. ![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/874395/image.jpeg)
![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/874395/image.jpeg)
0 Kommentare
Antworten (3)
Mahesh Taparia
am 8 Mär. 2022
Hi
With respect to your image, there are possibly 2 different colors. You can analyze the pixel values using histogram of image, refer this documentation for more detail. You can convert the RGB image to gray scale or you can consider the channel which you want to analyze. The histogram will give peaks, those peaks can be the desired value.
0 Kommentare
Mathew Metzger
am 30 Jul. 2022
I am also interested in color uniformity calculations. Have you found any solutions to your question?
1 Kommentar
Walter Roberson
am 30 Jul. 2022
You could find the standard deviation of some measure of the color. For example you could convert to HSV and examine the H channel -- remembering, though, that any difference from the mean needs to be calculated as min(difference, 1-difference) because H is an angle in the range 0 to 1
Image Analyst
am 30 Jul. 2022
Bearbeitet: Image Analyst
am 30 Jul. 2022
This is not hard at all. First of all you have to get just the green paper disc, which you can do easily by thresholding on the Saturation or CIE A color channel.
Then you need to convert the image to CIELAB with rgb2lab(). Then get the mean L, a, and b with mean(). Then you need to get a delta E image using Pythagorean theorem. Finally look at the histogram of the delta E values. You can use std or iqr to get a metric related to uniformity.
In short (untested)
labImage = rgb2lab(rgbImage);
[lImage, aImage, bImage] = imsplit(labImage);
% Compute the circular mask.
mask = aImage < -8; % Or whatever works. Or use the Color Thresholder app.
mask = imclearborder(mask);
mask = bwareafilt(mask, 1); % Take largest blob.
% Get mean LAB values.
meanL = mean(lImage(mask))
meanA = mean(aImage(mask))
meanB = mean(bImage(mask))
% Compute the color difference image.
deltaEImage = sqrt((lImage - meanL).^2 + (aImage - meanA).^2 + (bImage - meanB).^2);
deltaEImage(mask) = 0; % Mask away stuff outside the mask.
imshow(deltaEImage, []);
% Compute uniformity metrics:
stdDev = std(deltaEImage(mask))
If you can't get it to work, let me know and I'll work up a full demo.
0 Kommentare
Siehe auch
Kategorien
Mehr zu Modify Image Colors finden Sie in Help Center und File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!