Modeling in digital rock analysis: an image segmentation tutorial

Resumo

Image-based simulations are fundamental to assist in rock characterizations using a fully digital envi-
ronment. The basic workflow in digital rock physics or petrophysics comprises three general phases: 1) 3-D scan:

image acquisition and reconstruction, 2) 3-D modeling: image processing and segmentation, and 3) 3-D analysis:

image simulation and characterization. In this work, we focus in the modeling phase, by presenting two educa-
tional approaches to perform image segmentation of rock samples. The first approach relies completely in the use

of software with user interfaces, whereas the second one is a fully script-based approach relying in open source

packages with Python APIs. Although we describe educational approaches, which adopt compact sets of dedi-
cated strategies, they can also be combined to achieve the best performance and learning process. We assume here

that the images were acquired by high-resolution X-ray micro-computed tomography (micro-CT), and that they
will be employed to perform rock characterizations, such as determination of the effective absolute permeability
of porous materials based on the concepts of numerical homogenization. Therefore, step-by-step procedures and
implementations to achieve high quality segmented images are described.