How to use RGB transformation in to HSV and extract different layers and proccess it?

My task will be to map the intensity from the RGB model to the HSV model, where I will perform filtering. Then map the resulting image back to the RGB model.

Any tips or recommendation where I have to start and where I have to be careful in doing it.

I have the jpeg image which is 612 × 408 ×3, first i need to normalize image in to interval from 0-1, then use transformation function whit following parameteres: vi = max(ri , gi , bi) ci = max(ri , gi , bi) − min(ri , gi , bi) li = vi − ci 2 hi = 60deg · ( gi−bi ci ), vi = ri 60deg · (2 + bi−ri ci ), vi = gi 60deg · (4 + ri−gi ci ), vi = bi 0, ci = 0. (redi, greeni, bluei).

In the 3. part in need fold model in to proper fit:

The obtained values ​​of 'hi' , 'si' , 'vi' , which make up the HSV record (hue, saturation and value), fold in an array of the same size as the input RGB image so that each image element will have (xi , yi) values ​​(hi , si , vi).

Thank you.

Answer 1

[Image Analyst]

Use the built-in functions rgb2hsv and hsv2rgb

hsvImage = rgb2hsv(rgbImage);
[hi, si, vi] = imsplit(hsvImage);
% Then do filtering of any of them.
% Then rebuild a filtered RGB image
hsvImage = cat(3, hi, si, vi);
filteredRGBImage = hsv2rgb(hsvImage);

[Image Analyst]

To normalize you can use rescale() though I don't see why it would be necessary.

[Math]

Thank you for quick answer. My taks is to implement the code with my own idea, not with bulit-in function.

That is the reason why I aws askin for.

Thank you for understanding.

[Image Analyst]

Oh, okay. You forgot to tag it as homework so I'll do that for you. And since you're not supposed to use built-in functions, I imagine your professor would also not allow you to use any code that we gave you. I imagine he wants to see your code not mine, and you may even get into trouble for turning in someone else's code as your own. But good luck - it seems like you're using the right approach.

[Math]

something like this one function. Thank you for help and recommendations.

function oHSV = rgb2hsv(iRGB)

[Y, X, Z] = size(iRGB);

oHSV = zeros(Y, X, Z);

for y = 1:Y
    for x = 1:X
        R = iRGB(y,x,1);
        G = iRGB(y,x,2);
        B = iRGB(y,x,3);

        V = max([R,G,B]);
        C = V - min([R,G,B]);
        
        if V == R
            H = 60*(G-B)/C;
        elseif V == G
            H = 60*(2+(B-R)/C);
        elseif V == B
            H = 60*(4+(R-G)/C);
        elseif C == 0
            H = 0;
        end

        if V == 0
            S = 0;
        else
            S = C/V;
        end
        oHSV(y,x,1) = H;
        oHSV(y,x,2) = S;
        oHSV(y,x,3) = V;

    end
end

[Math]

Are you employed as instructor also? Is there any possibilities for it?

[Image Analyst]

I'd use imsplit and cat and do the whole thing vectorized, using the whole arrays rather than a nessted for loop. And of course I'd put in tons more comments like all professional programmers should do.

I do give internal short courses occasionally as part of my job for my employer but that's not my main job.