[feature] : Multiple rounds of traditional thresholding for simple multi class thresholding

[SylviaWhittle]

Is your feature request related to a problem?

No response

Describe the solution you would like.

As part of the plan for the grains refactor, we will need to be able to do simple multi class segmentation based on multiple traditional height thresholds.

This should be simple to do if we iterate through a list of thresholds provided by the user, forming multiple binary masks that we form together into a tensor.

Describe the alternatives you have considered.

N/A

Additional context

Should be done after the initial grains refactor.