Define right model and test design approach for colour measurements of plastic flakes

[01Mercy]

Hi all,

I am looking for the right model and right test desing to use for the following case:

I want to measure the colour of a flake material on a colour measuring device. There can be x (to be defined) different colours in the material. Standard material is blank.

Each colour I can measure in a dillution serie to make a calibration line for that specific colour y = ax + b

For multiple colours I than would use y = b0 + a1x1 + a2x2....

This approach however does not hold since within the measurement the different colours interfere with each other. I.e. I can make a perfect line with green and seperate one with red but if they are together in one sample they will give a yellow signal in y and my model will not hold.

So basically I would like to know what model I could use for this and what test design approach I can best use to feed the model with data.

Thanks a lot!

 

[IRstuff]

The question is what are you attempting to get out of the measurement? Your model does not work because that's not really how colors work. I suggest that you read the Wikipedia article to start with:

http://en.wikipedia.org/wiki/Color,

followed by the article on color space as currently used in most industries:

http://en.wikipedia.org/wiki/CIE_Color_space

and

http://en.wikipedia.org/wiki/Tristimulus_colorimeter.

The latter will lead to you to look at commercial colorimeters.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529