Flame temperature based on colour

[richard_z]

So I've been wondering, is it possible to determine how hot a flame is based on the colour? I know that you can classify it into general groups (yellow ~= 1000*C, etc.), but is there any way to get a more accurate temperature estimation? Any papers on this?

 

[Feldoh]

You'd need a device to very accurately measure the wavelengths, but I can't think of anything stopping you.

However I feel like there are more accurate and easy ways to measure temperature. Generally speaking I think it's more common to use color to analyze temperature distributions of a specific object.

 

[Vanadium 50]

I think you need to be more quantitative - i.e. measuring the continuum wavelength distribution and fitting it. Otherwise you can be fooled by emission lines - put a tiny bit of a sodium salt in a blue flame and it turns yellow.