How do I do a density profile? (Plot: Gray value VS. Distance, Area fraction VS. Distance ? )


I’m very new at ImageJ (Fiji) and I’m currently analyzing some scans from a micro-CT. My sample is cylinder-shaped, 20 mm in diameter and 3 mm thick, and I’m scanning the sample with the thickness in the z-direction.
I’m supposed to analyze the density profile (how the density differs through the sample thickness).

I think one way to do that is just to draw a line through the thickness and plot a line profile with Gray value vs. Distance(pixels). Another way is to measure the area fraction of each scan through the sample and plot Area fraction vs. Distance.

I think the last idea is the best, and I know how to measure the fraction for each scan in Fiji, but can I do some commando or something to automatically measure the fraction throughout the sample? And hopefully plot it as well.

The next problem is where do I start the measurement, my sample is a bit rough. Is there any function I can use that tells me where my sample top/bottom surfaces are and in which slices it is “good to start” with?

Good day,

from what you write I conclude that you are dealing with a stack of images and you need the indicated measures as a function of the slice number, i.e. discrete depths z. Could you please post the image of a typical slice (in PNG- or TIF-format), or, if the stack isn’t too big, perhaps a typical stack. In any case, it would be helpful to know the number of slices of a typical stack. You write that the thickness is 3mm, so what is the slice increment?



Thank you for your answer.
I just have a 3D picture of the reconstructed sample on this computer. It is a fiber material with very small fiber, so the slices look like the war of the ants.
Yes, my sample is around 3 mm thick and I import the stack to Fiji with increment 1, so the stack contains about 650-700 slices.

Thanks for the perspective view!

Because you prefer to obtain a plot of “Area Fraction” as a function of depth, we should like to know what you call “Area Fraction”. For this purpose it would be helpful to see a single and typical slice (in PNG- or TIF-format) for which you formally specify what you consider the “Area Fraction” of this slice.

From what I see, the slices aren’t binary-valued, hence the “Area Fraction” must be determined according to a criterion (threshold or else). We need to know this criterion, otherwise automatic evaluation is impossible.