[Soc-2005-cvs] CVS commit: soc-blender/sketching/include
sketch_type.h soc-blender/sketching/src main.cpp sketch_type.cpp
Pablo Diaz-Gutierrez
ihaveajob at gmail.com
Sat Aug 20 22:17:16 CEST 2005
pdg (Pablo Diaz-Gutierrez) 2005/08/20 22:17:16 CEST
Modified files: (Branch: pdg)
soc-blender/sketching/include sketch_type.h
soc-blender/sketching/src main.cpp sketch_type.cpp
Log:
Normal calculation for junction points with degree 3 or higher.
These normals are calculated as the normalized resultant of all the incoming segments
to the junctions of degree 3 and higher. A button has been added to flip normals in
case they're in the wrong direction.
There exist some obvious counter-examples to this approach, but we can either:
1.- Ignore them for the moment and focus on the cases when this works.
2.- Provide a rotation control to modify precalculated normals.
3.- Improve the normal calculation scheme to account for coherence of orientation with
respect to the neighbors. This would be the equivalent of a median (Laplacian?) filter
in the normals of the neighborhood of each junction point.
Anyway, so far we're going for the first solution, a.k.a. The Ostrich Approach (tm).
Next:
- Manual identification of faces. The user succesively marks the segments that
compound each face in counter-clowkwise order. Existing faces should be viewable,
selectable and erasable with the usual one-button approach.
Revision Changes Path
1.1.2.3 +10 -1 soc-blender/sketching/include/Attic/sketch_type.h
<http://projects.blender.org/viewcvs/viewcvs.cgi/soc-blender/sketching/include/Attic/sketch_type.h.diff?r1=1.1.2&r2=1.1.2.3&cvsroot=soc-2005&only_with_tag=pdg>
1.1.2.3 +28 -13 soc-blender/sketching/src/Attic/main.cpp
<http://projects.blender.org/viewcvs/viewcvs.cgi/soc-blender/sketching/src/Attic/main.cpp.diff?r1=1.1.2&r2=1.1.2.3&cvsroot=soc-2005&only_with_tag=pdg>
1.1.2.3 +215 -44 soc-blender/sketching/src/Attic/sketch_type.cpp
<http://projects.blender.org/viewcvs/viewcvs.cgi/soc-blender/sketching/src/Attic/sketch_type.cpp.diff?r1=1.1.2&r2=1.1.2.3&cvsroot=soc-2005&only_with_tag=pdg>
More information about the Soc-2005-cvs
mailing list