Hello everyone
I am working on geometric modeling of a core sample using CT scan and Materialise Mimics software. The emphasis is on creating a representation of holes and cavities inside the core. Due to limited access to industrial micro-CT, we were compelled to use a medical CT scanner and a GE Brightspeed 16 slice scanner was used for this purpose.
The core sample is a 37.5 mm diameter cylinder with a height of 45 mm. Moreover, a slice thickness of 0.2 mm was used during the scan and the pixel size is about 0.3 mm. I noticed that the Scanning Field of View (SFOV) of the scan was set to 150 mm. I presume the SFOV is very large for this case and a considerable area of the image scope is showing the blank area around the stone. Does anyone agree that if the SFOV was smaller, the quality of acquired images would have been better?
Also, I wanted to know what is minimum "detail", in other words, the minimum hole size that I can rely on using a CT scan with these parameters?
I've attached a sample of 2D view of the core.
Any recommendation is highly appreciated.
Thank you
Mohammad
It is a very interested question
I agree with Erik answer
Hi
Image-based geometric modelling and mesh generation” integrates image processing, geometric modelling and mesh generation with finite element method (FEM) to solve problems in computational biomedicine .
The “resolution” is ultimately what you are looking for and this isn’t entirely dependent on the FOV, but rather a combo of the chosen FOV and matrix size. For simplicity sake, say you have an FOV of 240 mm x 240 mm and you chose a matrix of 240 x 240, your in-plane resolution would be 1 mm x 1 mm. If you change the matrix to 480 x 480 then your resolution will then be 0.5 mm x 0.5 mm. The slice thickness will then determine the 3 dimensional voxel size.
The resolution depends upon the scanner model and detector size. I do not know the specs for your model, but we routinely acquire 0.5 mm resolution in humans. Since radiation dose is not an issue for you, you could probably do better.
Do you have access to micro-CT? This kind of scanner has clinical use for dentistry and has a much higher resolution than medical CT scanners.
Dear Erik Middlebrooks,
Thank you for your contribution.
Actually I do know the exact resolution of the voxels and the 0.3 mm pixel size that I mentioned in the question was calculated by dividing the SFOV of 150mm by the matrix size of 512 which is 0.293 mm that I mentioned as "about 0.3 mm ".
The part that is questionable for me is that with regards to the artifacts that are known in a CT scan (partial volume effects etc) that can make the holes in this core sample seem smaller (or maybe bigger) than what they are in reality, how far can the hole size in a CT scan can be accurate and reliable?
Thank you
Dear Bernardo Martins,
Thank you for your advice.
Unfortunately during the acquisition of the scans, we were unable to access a micro-ct scanner. But we are trying to get access to such scanner as its much high resolution will give considerably more accurate result.
Dear all,
I recently found a paper by Li et al which describes an interesting method for segmenting cracks and vugs of a core plug from a typical-resolution CT scan. Authors used the histogram of the greyscale values and curve fitted a normal distribution function to it. Then used the function's parameters to find a threshold for segmenting the vugs.
Here's the link to that paper:
Article Fracture and Vug Characterization and Carbonate Rock Type Au...
If anyone had done a similar approach, please send me a message.
Best Regards
- Badges
- Science topic