RGB image to HSV image conversion

Hello all

I have an rgb image and I want to convert it into HSV image and want to get only SV image.I have code like that

img=imread('lena.bmp');
[H S V]=rgb2hsv(img);

this computes each H,S and V channel individual I want to compute only SV only how can I do it..

it is not possible to o it like

cat(2,S,V);...?

or

rgbimage=img;
rgbimage(:,:,2)=S;
rgbimage(:,:,3)=V;

Any help is appreciated...

Answer 1

[Image Analyst]

No - none of that is right. For one, you can't get the separate channels out of rgb2hsv(), though that might be a nice new enhancement/feature to add. You can do this:

hsv = rgb2hsv(rgbImage);
h = hsv(:, :, 1); % Hue image.
s = hsv(:, :, 2); % Saturation image.
v = hsv(:, :, 3); % Value (intensity) image.

You certainly would not want to combine them into a 3D image called anything like rgbimage because that would be very deceptive since it's not an RGB image.

[Algorithms Analyst]

Yeah thats right.Ok I got the H,S,V image.Now I want to work only on SV channel not individual but combine. how can I combine them.How can I show SV image?

[Image Analyst]

I don't know what that means. You should work on them separately - how can you not and have it make any sense. For example if you combined them into a 3D matrix and did

 meanValue = mean(SV(:));

you'd get total nonsense. It only makes sense to me to work on them separately.

You can use imshow(), imshowpair(), or imfuse() to display a composite image.

[Algorithms Analyst]

ok. Now I got the S image.from above code.

fs=1.25; ts=0.8;

I want to show some good appearance of S image according to algorithm

if (Simage<ts)

S_out=Simage.*fs; end if Simage>=ts S_out=1; end

S_out is the same vector like Simage for storing output image.But the result is same.I got the same result.There is no change when I apply the above condition.Am I doing any mistake here?

[Jan]

It depends on what "combine" exactly needs. What do you think is an SV image? To display it, you have to choose any value for hue, e.g. 0 or 0.5. Otherwise the color is not defined unequivocally.

[azam sayeed]

sir when i apply meanValue to different images i still get the same mean value based on hue component

%window1
hsv2=rgb2hsv('testw1.png');
hueImage2 = hsv2(:,:, 1);
meanHue2 = mean2(hueImage2);
display(meanHue2 );

%window4
hsv3=rgb2hsv('test.png');
hueImage3 = hsv3(:,:, 1);
meanHue3 = mean2(hueImage3);
display(meanHue3);

output:

meanHue2 =

     38.9913

meanHue3 =

     38.9913

[Image Analyst]

You need to send in an image to rgb2hsv, not a string that happens to be a filename:

rgbImage = imread('testw1.png'); % Get image from file
hsv2=rgb2hsv(rgbImage); % Convert image, not string.