snow_partitioning_parallel

snow_partitioning_parallel(im, r_max=4, sigma=0.4, divs=2, overlap=None, cores=None)[source]

Performs SNOW algorithm in parallel (or serial) to reduce time (or memory usage) by geomertirc domain decomposition of large images.

Parameters
  • im (ndarray) – A binary image of porous media with ‘True’ values indicating phase of interest.

  • overlap (float (optional)) – The amount of overlap to apply between chunks. If not provided it will be estiamted using porespy.tools.estimate_overlap with mode='dt'.

  • divs (list or int) –

    Number of domains each axis will be divided. Options are:
    • scalar: it will be assigned to all axis.

    • list: each respective axis will be divided by its corresponding number in the list. For example [2, 3, 4] will divide z, y and x axis to 2, 3, and 4 respectively.

  • cores (int or None) – Number of cores that will be used to parallel process all domains. If None then all cores will be used but user can specify any integer values to control the memory usage. Setting value to 1 will effectively process the chunks in serial to minimize memory usage.

Returns

regions – Partitioned image of segmentated regions with unique labels. Each region correspond to pore body while intersection with other region correspond throat area.

Return type

ndarray