Author Topic: .  (Read 654 times)

0 Members and 1 Guest are viewing this topic.

Offline qwerty1Topic starter

  • Newbie
  • Posts: 1
  • Country: tr
.
« on: November 06, 2019, 07:29:30 am »
.
« Last Edit: November 10, 2019, 07:50:45 pm by qwerty1 »
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 7017
  • Country: pl
Re: Spectral sensor as color sensor
« Reply #1 on: November 06, 2019, 07:43:06 am »
You first need to decide which RGB standard you want to encode in, it's not like there is one. This decision has probably already been made for you by the system that's going to consume your data, whatever it is.

Then there is whole science of "color spaces" I'm not very familiar with. But the gist is that human vision has three types of cones with different sensitivity to different frequencies. So you somehow calculate how much each type of cone would be stimulated by the signals you detected and then you construct an equivalent stimulation using the particular RGB colors provided by your display.
 

Offline flynwill

  • Regular Contributor
  • *
  • Posts: 143
Re: Spectral sensor as color sensor
« Reply #2 on: November 06, 2019, 02:47:05 pm »
It depends on what you mean by "determine color precisely".   If the goal is to produce the same appearance to human viewers on a monitor of some sort, then you are looking for a sensor that can sense (or emulate) the CIE color matching functions.  You'll then need to apply the appropriate matrix to convert the resulting x,y,Y values these to whatever colorspace your display uses.

If your goal is something else then let us know.
 

Offline flynwill

  • Regular Contributor
  • *
  • Posts: 143
Re: Spectral sensor as color sensor
« Reply #3 on: November 07, 2019, 02:30:14 pm »
Ah, in that case a multi-band sensor like the one you located might work pretty well.  But you will likely need to calibrate it with actual test strips.  Your color chart reference will have been made with half-toned inks to give the same visual appearance as the actual test strip, but there will likely be significant spectral differences that a multichannel sensor will see differently.  The manufacturer of the test strips might have actual spectral response data on the dyes in the test strip that could also be the basis for working out a proper calibration process.  if your project has significant backing then the purchase of a real spectral-photometer might be in order, that would allow you to get the same information directly.

You will also need to pay close attention to the spectra of whatever light source is used to illuminate the sample.  Traditionally incandescent bulbs are nearly always used for this purpose because of their uniform spectra.  It might be possible to use a white LED light source for your project, but the very uneven spectra is going to complicate things.
« Last Edit: November 07, 2019, 02:36:15 pm by flynwill »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf