a comment: you will need to manage boundary cells accordingly too (else all of them be homogeneous Neumann), an alternative solution is to convert your data into standard input of deal ii mesh, e.g. UCD mesh format (its very easy for voxelized geometry), then simply load it by available functionality in deal ii. To do this it's suffice to take a look at "grid_in.h".
On Wed, Dec 8, 2010 at 5:30 PM, Wolfgang Bangerth <[email protected]>wrote: > > > For the example I gave above, which is 3x2x2, the example points are: > > .... > > ... > > 2 0 0 > > 3 0 0 > > 0 1 0 > > ... > > ... > > > > Each coordinate above represents a voxel. So, essentially, each voxel is > a > > cell !?! > > So let's assume that each voxel in your mesh has its front bottom left > corner > at the coordinates you give above (I don't know if that's what you want, > but > let's assume; it shouldn't be hard to make these points the center of the > cell, or multiply all of this by a certain factor Delta x to account for > the > proper size of voxels). For simplicity allow me to also forget about the > z-coordinate and assume that we are only dealing with a 2d mesh. Then the > first of your cells has vertices > 2 0 > 2 1 > 3 0 > 3 1 > The second of the cells above has vertices > 3 0 > 3 1 > 4 0 > 4 1 > And so on, and so forth. So you will want to create all of these vertices > in > one big array, and everytime you add one to the array you should check > whether you already have it. In the example above, the first two vertices > of > the second cell you already have, so in the end for these two cells, you > only > need the following six vertices: > 2 0 > 2 1 > 3 0 > 3 1 > 4 0 > 4 1 > At the same time as you add, you need to record for each cell which > vertices > they are composed of. For the first cell you will find the following: > 0 1 2 3 > because none of the vertices of this cell were already in the array. For > the > second cell you will find > 2 3 4 5 > This way, as you loop over your set of MRI points, you will build up a > global > list of vertices and a list of 4-tuples indicating the four vertices of > each > cell. > > Best > W. > > -- > ------------------------------------------------------------------------- > Wolfgang Bangerth email: [email protected] > www: http://www.math.tamu.edu/~bangerth/ > > _______________________________________________ > dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii >
_______________________________________________ dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
