GeoDict Forum

Image Processing and Image Analysis => CT, µCT and FIB-SEM => Topic started by: andrefeliperav on December 07, 2022, 02:34:21 PM

Title: Segmenting micro-CT cylindrical images / edges as porosity
Post by: andrefeliperav on December 07, 2022, 02:34:21 PM

On October 22nd 2019, we had a customer that was testing the software and complained about the possibility of using non-cubic (prismatic) voxels. At that version, GeoDict was only able to work with cubic voxels.


Currently, one of our customers said the user is facing a problem in segmenting micro-CT images in GeoDict, as the software is identifying the edges of the image (cylindrical) as porosity.Then, the user would like to check if there is any methodology to solve this issue, any tutorial that explains how to perform the segmentation correctly without the edges being considered as part of the porosity.


The user is working with a set of images that correspond to samples of very complex rocks in a cylindrical format. It would be crucial to be able to process them in GeoDict without having to cut them into a rectangle so as not to lose information. However, when using the original cylindrical format, the segmentation ends up confusing the edges of the image with porosity, as previously said.


Does GeoDict have a compatible solution?
Title: Re: Segmenting micro-CT cylindrical images / edges as porosity
Post by: Anne Blumer on December 12, 2022, 02:10:44 PM
Hi andrefeliperav,
GeoDict still works with only cubic voxels, but there is another way to cope with your struggles. The basic idea is to put some solid material around the cylindrical image part. To achieve this there exist two different ways in GeoDict with slightly different results. Depending on your further analysis you can choose which is best suitable for you. Both methods are described in detail in the PowerPoint Slides attached to this post.
I hope this will help you. If you have any more questions, please ask again.
Best regards,Anne