RGB image in the yellow green mixed color?

i have an RGB image and i separate it's three components R, G, B (image in Red component,.......) to see my image in each component ,i convert it from rgb2cmy and i can plot it in associted red and green components or green an bleu components ....but i can't do that for yellow and green associed components ?

i need to plot my image in associted two components yellow and green

thank you in advance for your help

 RGB =imread('image2.jpg');
 G = RGB;
 G(:,:,1) = 0;
 G(:,:,3) = 0;
 subplot(2,2,1);
image(G);
axis image
title('Green Component')
y=255-B;
subplot(2,2,2);
image(y);
axis image
title('yellow Component')
 GY=y+G;???????????????????
subplot(2,2,3);
image(GY);
axis image
title('yellow and Green Components')

[djamaleddine djeldjli]

I want my image to contain only the colors between yellow and green

Answer 1

[Image Analyst]

I'm not sure what you mean by the yellow component. The image has only red, green, and blue components. If you want you can use rgb2lab() and display the b component. Positive b would be yellow colors.

colorcloud(rgbImage, 'lab');
labImage = rgb2lab(rgbImage);
bImage = labImage(:,:,3);
figure;
imshow(bImage, []);

[djamaleddine djeldjli]

thank u Image Analyst

I want my image to contain only the colors between yellow and green

[Image Analyst]

Perhaps you should use the Color Thresholder and do it in HSV color space. Set the hue to contain only colors in that hue range.

[Allinone]

Is it possible to see two colours in an RGB image and then convert it in a gray scale?

[Image Analyst]

You can have two, or millions, or any number of colors in an RGB image. To visualize/see them, use the imshow() function:

imshow(rgbImage);

Then, to convert the RGB image into grayscale, use the rgb2gray() function:

grayImage = rgb2gray(rgbImage);

Answer 2

[Guillaume]

"I want my image to contain only the colors between yellow and green"

First, you need to define mathematically what that means.

As IA says, you're probably better off working in the HSV colour space. You can easily set a threshold on the hue for whatever you define as between yellow and green (maybe 52/360 to 154/360 ?). Even then, you probably also need thresholds for value and saturation. For example:

yellow1 = hsv2rgb(cat(3, 57/360, 1, 1));  %yellow at full saturation and brightness
yellow2 = hsv2rgb(cat(3, 57/360, 1, 0.1);  %same yellow at full saturation but much lower brightness
imshowpair(repmat(yellow1, 200, 200), repmat(yellow2, 200, 200), 'montage');

The 2 images generated above are exactly the same yellow hue. Yet not many people would say that the right hand image is yellow.

So maybe:

rgbimage = hsv2rgb(cat(3, mod(hankel(50:249, 249:448)/200, 1), repmat(1, 200, 200), 1-toeplitz(1:200)/200));  %demo image
hsvimage = rgb2hsv(rgbimage);
isyellowgreen = hsvimage(:, :, 1) >= 52/360 & hsvimage(:, :, 1) <= 154/360 & ... filter on hue, adjust as required
                hsvimage(:, :, 2) >= 0.2 & ... filter on saturation, adjust as required
                hsvimage(:, :, 3) >= 0.2;    % filter on value, adjust as required
filteredimage = hsvimage;
filteredimage(repmat(~isyellowgreen, 1, 1, 3)) = 0; %set non yellow-green to 0
imshowpair(rgbimage, hsv2rgb(filteredimage), 'montage')

Adjust thresholds as required

[djamaleddine djeldjli]

thank for your answer what i need is for example : i have an image of a landscape dans i want it in a component which represents a red to green color indicator so i do :

 RGB =imread('image2.jpg');
 luvimg=rgb2luv (RGB);  % just to see conversion 
 imshow (u)      %  U component represents  red to green color indicators

so what i need is a comonent which represents yellow to green color indicators

[Guillaume]

In which way my answer from So maybe does not do what you want?

[Image Analyst]

djamaleddine, I think you need to learn more about color space. Red and green are complementary colors. That's why they show up on opposite sides of the color gamut. See my CIE LAB icon in my answer. Likewise, blue and yellow are complementary. That's why you can take the b component of LAB as a measurement of yellowness to blueness, and the "a" component of LAB as a measure of the greenness to redness of a signal. Just look at any 3-D color gamut. Now, you can get green pixels and yellow pixels by thresholding in the hue domain but to get the yellowness to greenness amount, you'd have to rotate the LAB axes by about 45 degrees. It's very unconventional and I've never heard of it being done or of anyone ever needing it. Even it's meaning/interpretation is confusing since green and yellow are not complementary. The big question is WHY do you think you need to do this?

[djamaleddine djeldjli]

Image Analyst , Guillaume thank you for your interest

I am doing this to measure the absorption of hemoglobin in a bowl of blood. So hemoglobin absorbs light between 540 nm and 577 nm wavelengths corresponding respectively to green and yellow colors. try to see the picture that contains my results to understand me

as you see in the photo I can combine two colors eg blue and green but I can not do it for yellow and green

[Image Analyst]

I still don't think your approach is right, but whatever...let's discuss... Since yellow is red + green, if you want green + yellow you'd want green + green + red. However that double counts the pixels' contribution from the green wavelength part of the spectra, so you need to subtract that out. So if you want yellow + green you'd need to subtract out the red. However red contributes to yellow. There is no yellow in RGB images without red. But if you subtract out the red, you're just left with green. You see, digital cameras have 3 spectral sensitivity ranges, each taking up about a third of the spectrum but with a lot of overlap. There is no sensor that just looks at yellow, only green and red. So I'm back to the original suggestion which is if you want pixels in the yellow to green range, then you should do color segmentation on the hue channel in HSV color space. And if you want the amount of Yellowness/blueness in any color, then you should convert to LAB color space and use the b channel.

[Guillaume]

I don't really understand the question anymore. As Image Analyst just explained there isn't a yellow channel in RGB (which is what matters for display), so you can't isolate yellow in RGB. It's a mix of red and green.

You can isolate yellow (and green) in the HSV colour space, which is what I've shown in my answer.