Ice Core CT scan, threshold results in imaginary surface of ice that lines bubbles in resin used to store the core.

Hello,

I am analyzing images of ice core CT scans and the analysis of the core is skewed due to imaginary surfaces of ice in bubbles that are inside a resin used to store the ice.

the lightest gray is the resin, the darker is ice, and the black is air. the "imaginary" ice forms between the black and light gray because the gradient at that boundary has values that are counted as ice when thresholded. I have found a solution that involves identifying the areas of highest gradient and then forcing them to be deffer to the neighboring materials at a 50/50 split. I know how to identify the areas of high gradient but i do not know how to make the areas go to the proper material. Is this a proper method to analyze this or is there another more common method?

Answer 1

[Image Analyst]

Take the gradient image that identifies the edges. I don't know what "deffer" means, but I assume that the edges (dark/ice) are what you want to replace with resin/light. So just do this:

gradientImage = imgradient(grayImage............whatever....
mask = gradientImage > someThreshold; % Identify edges.
grayImage(mask) = resinGrayLevel; % Replace edges with the resin's gray level