Graphics Reference
In-Depth Information
Figure 9.18.
Normal curvatures at a point.
n p
X u
p
u
S
Euler's Theorem. As u ranges over all unit vectors in the tangent plane to S at p ,
the set of curvatures k u , called normal curvatures , forms a closed interval [k 1 ,k 2 ]. If
the k u are not all equal, then there is precisely one direction u 1 for which the associ-
ated curve has the minimum curvature k 1 and one direction u 2 for which the associ-
ated curve has the maximum curvature k 2 . Furthermore, u 1 and u 2 are orthogonal
and if a unit vector u makes an angle q with u 1 , then
2
2
kk k
u =
cos
q
+
sin
q
.
(9.41)
1
2
Definition. The curvatures k 1 and k 2 are called the principal normal curvatures of S
at p . and the radii of the associated osculating circles are the principal radii . The
vectors u 1 and u 2 are called the principal normal directions of S at p .
Note that if we change the direction of the normal n p , then the sign of the cur-
vatures k u changes, but all the results remain valid. Another interesting fact is that
we really did not have to consider planar curves (curves that lie in X u ), but any para-
metric curve in S through p that has tangent vector equal to u would give rise to the
same curvature k u .
Let c u denote the center of curvature at the point p of the planar curve that was
used to define the normal curvature k u . The centers c u lie on the line normal to S at
p , so that c u = p + t u n p for some t u ΠR . The set of these centers as u ranges over all
unit vectors in the tangent plane consists of one or two closed intervals. More pre-
cisely, there exist numbers
t
and
t
,
-•£
t
£
t
£•
,
min
max
min
max
so that the set of centers is either
[
]
[
] »•
[
]
cc
,
or
-•
,
c
c
,
,
min
max
min
max
where c min = p + t min n p and c max = p + t max n p . The points c min and c max are the centers
of circles that define the principal normal curvatures k 1 and k 2 , respectively. See Figure
9.19. The point p in Figure 9.19(a) is called an elliptic point of the surface because the
surface lies to one side of the tangent plane and meets the tangent plane in a single
point locally. The point p in Figure 9.19(b) is called a hyperbolic point because the
Search WWH ::




Custom Search