Hi, I am currently working on a numerical simulation of the flow of two fluids through porous digital rocks using C++. I have a binary file(1=rock,0=void) for the rock which has a 2D slice-based structure meaning that the data representing 10x10x10 rock will have 10 columns and 100 rows with every ten rows representing a slice of a 3D structure. And also the data on a separate file with the same structure representing the position of the fluid in the voids(0's) of that rock. I have already stacked the 2D array to 3D using python but I don't know how to plot this as voxels to visualize the 3D rock and the fluid inside of it. I tried 3D voxel plots using matplotlib but it turned to be so slow for a large chunk of data(1000x1000x1000). It does not seem to be a tough challenge, however, I cannot find a proper way to do it. All I want is a 3D voxel plot with some transparency adjustments which is not very slow. For example, I'd like to plot a 3D binary array with 0's as void and 1's as solid and then on that figure the fluid with 0's as no fluid and 1's as fluid(similar to " hold on" functionality in matlab). It would be even nicer to give colors corresponding to the density of the fluid at each voxel since I have a array with float values as well. The file format which I write out from the simulation is just simple .dat format with containing space-separated float(or binary) values. Then from that file I just read the values as a 3D array using python. Below is a nice example of what I want to do.
https://ibb.co/RN7qQwn
It looks like you might want something like a marching cubes, naive surface nets, or dual contouring approach. There are a lot of resources on this, primarily in reference to video game terrain engines or biomedical imaging, but to get you started I would suggest this article: https://0fps.net/2012/07/12/smooth-voxel-terrain-part-2/
For a nice visualization of marching cubes, check out this video: https://www.youtube.com/watch?v=M3iI2l0ltbE
Your data size (1000^3) is also large, so you might have to split that up into chunks. Transparency and other effects can be stored as vertex data.
All of what you are wanting is very doable, and not too complicated if you have the prerequisite skills, but if you are starting from scratch (both in terms of code and understanding), it may take a bit. I've seen both Python and C++ code to do some of the things you are looking for, but you will likely have to spend some time adapting the code and your data structure.
Good luck.
Dear Orkhan Shirinov,
Look the link, maybe useful.
https://towardsdatascience.com/the-art-of-effective-visualization-of-multi-dimensional-data-6c7202990c57
https://www.quora.com/How-do-I-visualize-multidimensional-arrays
Regards,
Shafagat
In my opinion refer this link getting some ideas
https://towardsdatascience.com/the-art-of-effective-visualization-of-multi-dimensional-data-6c7202990c57
Thanks a lot for all of the recommendations ! I am now using a mixture of methods to visualize arrays in the simulation; OpenCV for creating 2D slices of the fluid in ".png" format and FEI Avizo to stack the images and create a 3D representation of the domain. It may be an unconventional approach but it does what I want and most importantly, simple and fast!
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python