Hello,
I have micro-CT reconstructed slices of my open cell iron foams made by NRecon software. I have an impression that the quality of the images is not as good as it should, and the struts tend to be blurry. This was reflected when I created the 3D model through the 2D slices in another software. I have been suggested to use the smoothing function in NRecon software or find a software which turn the grey images into the black and white slices and then use those black and white slice images to reconstruct the 3D model. Does anybody know such a software? Do you have any other suggestions? I would appreciate your advice!
Thanks,
Reza
Hi - If you are using Nrecon with no filtering the ram-lak default filter will amplify image noise in the tomogram. You could add a 1-2 pixel smoothing filter to reduce that or switch the reconstruction filter to a Hamming filter. In regard to your 3D models - all voxel based images without anti-aliasing features will look rough and distorted. If you are using a model building program that allows vertex smoothing on the model mesh you will find your models will improve. Such a function exists in meshlab an open source project. It is available at http://www.meshlab.net/ I hope this helps with your issue.
I recommend that you use free imageJ software his website
https://imagej.nih.gov/ij/
Reza Alavi can you show us some example images of:
a) your radiographs
b) your scan parameters
c) your reconstructed slices
Are you certain you have a quality scan; sufficient penetration, no movement, appropriate spot size for your resolution and feature sizes etc. Have you corrected properly for beam hardening? Can you apply a pre-reconstruction median filter.. this often helps much more than it hurts.
We offer a post reconstruction software to suppress noise and enhance resolution of CT images. In addition we offer high quality CT reconstruction software if you have access to the projection data. Feel free to contact me if you are interested.
The CLAHE script can improve contrast for image stacks, before you import the images into a reconstruction program: https://imagej.net/Enhance_Local_Contrast_(CLAHE)
Update: I have realized that changing the Misalignment Compensation value before starting the reconstruction can significantly improve the quality of the reconstructed slices (this was also mentioned in NRecon manual). It helped me to significantly reduce the unwanted "tails" which had been appearing.
Thanks a lot for sharing your answers. I think the scanning parameters are also important. How to use the proper voltage and adjust the source and detector distance all count for the final resolution of image.
- Badges
- Science topic