Quantitative analysis of the colour gradient

I am doing an experiment with a thermal camera. It produces a colour map of an image with a temperature scale, where each colour corresponds to a certain temperature. I need to quantitatively analyze the colour gradient to determine the temperature of a certain pixel on the image.

Here is the link to the image: https://drive.google.com/file/d/1IlM5BHrv-Nf1UOTeZIeA_M9R4WJX-E-o/view?usp=sharing
(Unfortunately, image.sc wouldn’t allow me to upload the image)

1 Like

Hello mrvanalog,
The pixel you noted is within a 27 x 27 Pixel square found using the wand tool set at 150 (left click and set) then Edit > copy, from there File > new image-32bit,27x27 paste finally Image > Adjust > brightness /contrast and you have your close up image. The gradient of the new image will be greater than the original due to the close up so you will have to apply your own LUT (color map) and the line tool with plot profile to get the numbers.


Also, sorry,Thermal
Bob again

Hello mrvanolog -

The best would be if you could get a “raw” image from your
thermal camera that does not have the false-color mapping

Second best would be if you could get the false-color mapping
from a spec sheet for the camera.

Last would require some programming, but you could read the
color scale bar to the right of the actual image directly from the
image pixels. You would then use that information to translate
the false-color pixel values into to thermal values.

I’m not sure exactly what you want, but if you literally want the
gradient of the actual temperature, once you get the scalar
thermal values, you would calculate their gradient.

Lastly, for convenience of the forum, I am posting a .png version
of the google-drive image you linked to:

Thanks, mm

1 Like