While poking around I came a cross some interesting papers that I thought I would point out:
Campen, M. & Kobbelt, L. Polygonal Boundary Evaluation of Minkowski Sums and Swept Volumes Computer Graphics Forum, 2010, 29, 1613-1622 Bommes, D. & Kobbelt, L. Accurate computation of geodesic distance fields for polygonal curves on triangle meshes Proc. of Vision, Modeling, and Visualization (VMV), 2007, 151-160 Bischoff, S. & Kobbelt, L. Parameterization-free active contour models with topology control The Visual Computer, Springer, 2004, 20, 217-228 hope this is interesting... EBo -- On Wed, 5 Jan 2011 07:12:01 -0500, Mark wrote: > On Jan 5, 2011 6:09 AM, "Anders Wallin" <[email protected]> > wrote: >> >> > There are a few other people working on the problem of modeling >> material >> > removal. Unfortunately it seems that our approaches are different >> enough >> > that merging the projects would be difficult. >> >> hi all, >> >> I've done a bit of work modeling the stock as a signed-distance >> field. >> I'm using the classic marching-cubes algorithm to extract the >> iso-surface triangles of the stock model. There are more advanced >> approaches which also store surface normals in each octree vertex. >> To >> save space the distance field is not stored on a regular 3D grid, >> but >> in an octree which is subdivided whenever required. >> the approach is described here: >> http://www.cadanda.com/V2Nos1to4_11.pdf >> http://dx.doi.org/10.1080/00207540410001671651 >> http://dx.doi.org/10.1115/1.3130231 >> >> my results are here: >> http://www.youtube.com/watch?v=1DAvgLCj_RQ >> http://www.youtube.com/watch?v=17mpgQ4KVLg >> >> The code is here: >> > > http://code.google.com/p/opencamlib/source/browse/#svn%2Ftrunk%2Fsrc%2Fcutsim >> >> This approach can deal with any kind of cutters as long as you come >> up >> with the distance field for the cutter, essentially a function that >> takes an xyz-position and returns positive outside the cutter and >> negative inside. >> To model cutting moves or sweeps the sampling-approach is used. The >> low-level operation (runs in 1ms or so) is to subtract the cutter >> positioned at some xyz from the stock. Linear/arc/etc. moves are >> dealt >> with by positioning the cutter at many sampled points along the path >> and calling the basic function which subtracts the (stationary) >> cutter >> for each sample point. >> >> I hope to do some more work on this in February. The to-do list >> could >> look something like this: >> - find a g-code interpreter (emc2 or other) which can be used to >> read >> g-code and output 'canonical' G1 type moves to the sim > > If you want one c++ object for each g00/g01/g02/g03, you may be able > to use > some of my code. > >> - work on accelerating the opengl 3D view. I hear this involves >> things >> like VAOs and/or VBOs which I haven't tried. >> - I'm now calling the lib/sim from python, but this is probably too >> slow. Think about a C++ GUI that would integrate >> interp+3D-view+cutsim >> - fancy sim improvements: >> -- different colors for stock cut with different tools >> -- material removal rate calculation/plot/warning >> -- collision detection (fixtures=forbidden stock region, >> toolholder=forbidden cutter region) >> -- improved marching-cubes: dual-contouring, hermite-data, etc. >> >> enjoy, >> Anders >> > > I need to figure out how to run cutsim! > Mark ------------------------------------------------------------------------------ Learn how Oracle Real Application Clusters (RAC) One Node allows customers to consolidate database storage, standardize their database environment, and, should the need arise, upgrade to a full multi-node Oracle RAC database without downtime or disruption http://p.sf.net/sfu/oracle-sfdevnl _______________________________________________ Emc-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
