What is it about?
The work is about a key step in Digital Rock where image processing is used to generate a 3D Digital Rock model from an image of rock taken by X-ray computed micro-tomography. We propose a workflow how to check which image processing routine gives the best results in terms of segmentation, i.e. deciding what is pore and what is grain in a grey scale micro-CT image. The workflow involves a "ground truth" image to which the results of the image processing are compared. This direct comparison allows to quantitatively assess the success of the image processing.
Featured Image
Why is it important?
Validating image processing is of key importance because the quality of all further properties computed from the digital rock model depend on the quality of the image processing. So far all validation attempts of image processing workflows are somewhat subjective or qualitative at best because no 1:1 ground truth was available. Our workflow allows to assess image processing workflows quantitatively, and in an automated fashion. That will allow progress in image processing much quicker and more accurately.
Perspectives
The idea of this work was born during the Christmas break 2017/2018. The tools used are to a large extent OpenSource which allowed me to get to the first results, which I then shared with my colleagues in Houston and Bangalore, who contributed strongly to the paper with different image processing methods that have been developed and implemented earlier.
Steffen Berg
Royal Dutch Shell plc
Read the Original
This page is a summary of: Generation of ground truth images to validate micro-CT image-processing pipelines, The Leading Edge, June 2018, Society of Exploration Geophysicists,
DOI: 10.1190/tle37060412.1.
You can read the full text:
Resources
Contributors
The following have contributed to this page
