Extract Red Line From JPEG

6 次查看（过去 30 天）
Samuel Leeney
Samuel Leeney 2021-2-17
回答： Image Analyst 2021-2-18
Hi There,
I need to extract this red line from the image.
The line was drawn on and the file saved as a JPEG so it may be possible to simply extract this element from the image.
I do not know the exact tone of red so am unable to simply search for pixels in that colour.
Is there a way I can detect the red parts of the image so that I can isolate them to use later as required?
I have tried using getpts to click on the red line however for some reason it always returns a different colour value.
clear all
im = imread('Images/Annotated/2x_LN063B1_an.jpg');
im_original = im;
im = imresize(im,[3200,3200]);%resize
imshow(im)
hold on
[x,y] = getpts;
x = round(x);
y = round(y);
imcolour = im(x,y)
  2 个评论
Star Strider
Star Strider 2021-2-17
Attaching the image could be helpful.
Samuel Leeney
Samuel Leeney 2021-2-17
Apologies, I thought I had!
I have added it now.

请先登录，再进行评论。

请先登录，再回答此问题。

采纳的回答

James Li
James Li 2021-2-18
Hi,
First of all, here is some general advice that might help you work on images:
1) To find the RGB value of a pixel, in addition to running getpts command in the prompt, you can also use the "Data Tips" tool from the Figure which might be easier. You can find it in the Menu - Tools - Data Tips.
2) It is typical to see different values for pixels that seem to have the "same color" -- JPEG is a lossy image format, so some values are altered during compression when the image was saved. And there could be other artifacts introduced in various stages of image processing that would lead to this kind of fluctuation in values.
One way to solve your problem is to define a range of the RGB values, i.e., a "standard" value for the red color and experiment with "tolerance", and use them to filter this red line.
However, for your specific example, since your background is on the grey scale, while the only element with color is the red line, you could convert your image to CIELAB (or L*a*b*) space and easily get the result you need. You can read more about CIELAB color space here: https://en.wikipedia.org/wiki/CIELAB_color_space.
I have prepared a script that demonstrates this process:
img = imread('image.jpeg');
subplot(1, 3, 1);
imshow(img);
title('original image');
lab = rgb2lab(img); % convert image to CIELAB color space
subplot(1, 3, 2);
imagesc(squeeze(lab(:, :, 2)));
title('a* value');
colormap gray;
threshold = 2; % arbitrary value
map = squeeze(lab(:, :, 2) > threshold);
subplot(1, 3, 3);
imagesc(map)
title('a* value after threshold');
colormap gray;
indices = find(map);
[rol, col] = ind2sub(size(img, 1), indices);
The general idea is to convert the image to CIELAB color space and look at its a* (or b*, it works similarly) value. Most of the background is near 0, while the area covered by the red line has a non-zero value. Then you can define a threshold and do some filtering to get a map of only 1 and 0 values. You can experiment with that threshold and see how it changes the result. I put some plotting code in the script to help visualize the process and attached the figure to my answer.
You can also then get the indices or these pixels by using find and ind2sub functions, as in the last two lines of the script.
Alternatively, you could convert the image to HSV color space using rgb2hsv and look at the image's H (hue) or S (saturation) channel, and you will get similar results.

更多回答（1 个）

Image Analyst
Image Analyst 2021-2-18
Try the Color Thresholder on the Apps tab of the tool ribbon. From there you can export code for a function that you can then call and use in your main program.

类别

在 Help Center 和 File Exchange 中查找有关 Image Thresholding 的更多信息

标签

产品

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by