tsx programming: geometry question 2004-04-11 - By Richard R. "Remnar"
Back
Hi, sorry if this is the wrong place to ask. I have been busy for days
adding a new feature into Scatter that will encapsulate a particle onto a
face of another polyhedron. What I need help with is analysis on data that I
extracted using a test program I wrote. I have tried and tried and tried to
use a bounding box to get all my info, but because of point-face
orientation, the bounding box never gave me the right results unless I
figure out how to create a rotation matrix from the 4 points and the face
normal and then invert it then transform the bbox so it is normalized along
the x-y plane. Since I really dont know how to do that yet, I sought out
another possible way and that is by calculating apothems (is that the right
word?) on triangles and Ngons. An apothem is a line from a vertex to the mid
point of the oposite site. Calculating this was easy, even for a N gon. Now
I got the data, of the size of the face, including the apothems, mid point
and length point and line lengths. however it came with 2 or more sets of
data that I do not know how to filter out and I have ruled out every
possible way I know. I am using a sphere to get this data because it has
several axii to test with (actually i should have used a torus to test
with). I have determined what the sizes i needed are by use of a point edit
selection on each face and finding out what section and face index each one
belongs to.
Sphere Face Index = 0 section 1
Location: 0.2454, 0.0488, 0.9493
points[0]: 0.0000, 0.0000, 1.0000
points[1]: 0.3536, 0.1464, 0.9239
points[2]: 0.3827, 0.0000, 0.9239
aperatures[0]: 0.2218 lengths[0]: 0.3902
aperatures[1]: 0.3830 lengths[1]: 0.1493 *
aperatures[2]: 0.2218 lengths[2]: 0.3902
target size section 1, 8 = 0.149, 0.383, 0
target size section 2, 7 = 0.276, 0.385, 0
target size section 3, 6 = 0.360, 0.388, 0
target size section 4, 5 = 0.390, 0.390, 0
A sphere has 8 sections top to bottom with faces of certain sizes. Each
section has faces = how many latitudes of the sphere.
What I want to find out is why out of the 3 aperature(maybe i got this word
wrong, it means apothem length) and length is the 2nd one the correct one? I
know its the correct one cuz i selected that face to find out the targe size
of the face and section. But the program doesnt know if it is or not.
Now here is 2 more data examples for a 4 sided ngon face on 2 different
sections but along the same longitude.
Sphere Face Index = 16 section 2
Location: 0.5242, 0.1043, 0.8155
points[0]: 0.3827, 0.0000, 0.9239
points[1]: 0.3536, 0.1464, 0.9239
points[2]: 0.6533, 0.2706, 0.7071
points[3]: 0.7071, 0.0000, 0.7071
congruant aperature sums[0]: 0.3850, Lengths 0.1493
congruant aperature sums[1]: 0.2126, Lengths 0.3902
congruant aperature sums[2]: 0.3850, Lengths 0.2759 *
congruant aperature sums[3]: 0.2126, Lengths 0.3902
Sphere Face Index = 64 Section 5
Location: 0.9253, 0.1841, -0.1913
points[0]: 1.0000, 0.0000, 0.0000
points[1]: 0.9239, 0.3827, 0.0000
points[2]: 0.8536, 0.3536, -0.3827
points[3]: 0.9239, 0.0000, -0.3827
congruant aperature sums[0]: 0.3899, Lengths 0.3902 *
congruant aperature sums[1]: 0.3753, Lengths 0.3902
congruant aperature sums[2]: 0.3899, Lengths 0.3605
congruant aperature sums[3]: 0.3753, Lengths 0.3902
You see how the first set there (asterisk) is the 3rd element that holds the
correct data, but now we have the last one here that is the first element. I
need help determining how to filter out the rest of the elements so I only
end up with the correct one. I have tried using the location to determine
what coordinate area the face is in. However, i dont think that will work in
all cases.
If you need the test code, well its all in Delphi 6 Enterprise.about 81
lines worth. I have built an alternative to all this math stuff and that is
to build a face with the vertex data and extract what i need from it.
However I prefer not to do this because it would be much slower than the
math version. If there is another plugin forum/mailing list I can onto about
tsx development, please let me know.
Thanks,
Richard R.
|
|
