4
$\begingroup$

I understand I can get a color from wavelength but it's not true for the opposite direction. In case this picture would help.

enter image description here

To make this question simple, let's lighting and calibration aside to an optimized condition.

I wonder if there is any way to get intensity value of a certain spectral wavelength for given pixel of a common image. That is, Wavelength and rgb of a pixel are both known. Based on that, I want to compute intensity of the wavelength for the pixel.

$\endgroup$
3
  • $\begingroup$ Well, I mentioned about what you are concerned in the first sentence. I am not looking for a inverted function. As I already have a value of wavelength as well as color (hue) of the pixel, I want to figure out how I can get intensity of the pixel for the wavelength. $\endgroup$ Commented Feb 18, 2012 at 17:29
  • $\begingroup$ Let me try this again to see if I understand: you have an RGB triple that indicates a color of a pixel. You also have a particular wavelength of light. You would like to know for an arbitrary RGB color how much content there is at the wavelength of interest? Seems like you would model an RGB color as the sum of three "pure" red, green, and blue components. You would need to make an assumption on what the wavelengths of those pure colors were, but for an ideal RGB color, there wouldn't be any other components in its spectrum. $\endgroup$
    – Jason R
    Commented Feb 23, 2012 at 14:10
  • $\begingroup$ @JasonR You understood correctly. I don't think I can achieve my goal with RGB. I hope I can convert it to something like CIE. And then I hope I will be on right track. Thanks for your comment. $\endgroup$ Commented Feb 24, 2012 at 1:12

1 Answer 1

1
$\begingroup$

If I understand correctly - you want to know what wavelength a given RGB value will show?

Only some RGB values represent a pure wavelength, those on the edge of CIE diagram (there is no wavelength of brown for example) see http://www.fourmilab.ch/documents/specrend/ for details and code

$\endgroup$
1
  • $\begingroup$ Nope that's not what I asking. Wavelength and rgb of a pixel are both known, I want to compute intensity of the wavelength for the pixel. But you are the second person who misunderstood my question. So fault is on me. I will update my question. Thanks for your answer anyway. $\endgroup$ Commented Feb 23, 2012 at 1:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.