How can I compare two separate images (before and after) and count how many pixels are in a particular spectrum?

2 次查看(过去 30 天)
I would like to quantify the number of pixels that are in a particular color range (glowing) in an image. There are two images below. I am interested in knowing how many pixels are flourescing on the subject's body. Image #1 should have zero (or close to zero) pixels glowing. Image #2 should have a great deal more.
  1 个评论
Walter Roberson
Walter Roberson 2020-5-26
I would tend to think of converting to HSV, selecting a range of hue, in combination with a minimum saturation.
My second thought was that you might be able to threshold on brightness, but then I looked again and saw that their are a couple of lights that might turn out to be bright even though they are not the color you are looking for, so I think my first thought of HSV would be a better approach.

请先登录,再进行评论。

回答(2 个)

Image Analyst
Image Analyst 2020-5-26
Try the Color Thresholder onthe Apps tab of the tool ribbon. Try HSV color space.

Walter Roberson
Walter Roberson 2020-6-2
Example:
RGB=imread('image.jpeg');
figure(1)
imshow(RGB)
HSV = rgb2hsv(RGB);
H = HSV(:,:,1);
R = RGB(:,:,1); G = RGB(:,:,2); B = RGB(:,:,3);
maskH = H < 0.4 | H > 0.6;
R(maskH) = 0; G(maskH) = 0; B(maskH) = 0;
rec = cat(3,R,G,B);
figure(2);
imshow(rec)
This might not be good enough for your purposes, as it misses some of the streaks on the floor. But the 0.4 and 0.6 were chosen arbitrarily so there is room to tune them. If you
figure(3)
imshow(HSV);
then you can use a data cursor to explore the HSV values of the pixels of interest.

Community Treasure Hunt

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

Start Hunting!

Translated by