Pore classification

[bbe020]

Hi,
I have a binary image of grains and a connected pore. and I was wondering if it's possible to classify the pores as pore throats and pores using the PoreSpy ?

[jgostick]

Definitely, but the best approach depends on what you want to do with the result. You can extract a pore network usin porespy.networks.snow2, then import it into OpenPNM. You can view OpenPNM networks in Paraview, and even overlay the original image and the network. Here is an example outline how it works. If you just want to divide the image into pores and throats, then maybe porespy.filters.local_thickness would be of interest. This will draw spheres in the image so idenfies pore bodies, though not throats. If you let me know more specifically what you're looking for I can suggest something.

[bbe020]

Hi,
Thanks for the suggestion. I want something more like the code described in this paper:
"Wei, B., Hou, J., Lei, Z. et al. A hybrid method to segment the pores and throats of micromodels. Arab J Geosci 12, 716 (2019). https://doi.org/10.1007/s12517-019-4925-8".

I'am running a foam experiment on micromodel, and I want to count the appearance of bubbles in each pore.
I don't think that porespy.filters.local_thickness will work for that purpose, as it doesn't define the pores properly.

The porespy.networks.snow2 can work, but it divides many pores to two or three pores.

[jgostick]

The definition of where one pore stops and another one begins is quite tricky. The snow function uses a watershed, and actually tries to reduce the amount of oversegmentation. You can adjust how much oversegmentation you get by changing the sigma and r_max arguments maybe. Check out the detailed example of calling the individual steps in snow here. You can also trim nearby peaks which might reduce the splitting of pores further.