COMPUTER AIDED MANUFACTURING

Curves and Surfaces

1
CURVES
2
 CURVES
• Described by an array of points.
• Required in many applications:
- model real world objects (CAD)
- fonts
- animation

polylines Smooth curve

3
 MATHEMATICAL REPRESENTATION OF CURVES
• The general form of a polynomial function is:

where 𝑛 is a nonnegative integer and 𝑎0, 𝑎1, … , 𝑎𝑛 are real numbers.

• The polynomial 𝑝 𝑥 is of degree 𝑛 and 𝑎𝑛 ≠ 0.

4
 IMPLICIT CURVES/SURFACES
F(x, y) = 0 2D-Planar curve

𝑆1 𝑥, 𝑦, 𝑧 = 0 3D- spatial curve

intersection of two
𝑆2 𝑥, 𝑦, 𝑧 = 0 surface equations

http://www.robots.ox.ac.uk/~ian/Teaching/CompGeom/lec3.pdf 5
 IMPLICIT CURVES/SURFACES

x2 + y2 − r 2 = 0

• Dr. Scott Schaefer

6
SIMPLE IMPLICIT SURFACES
• Spheres
• Planes
• Cylinders
• Cones
• Torus

• Dr. Scott Schaefer

7
IMPLICIT CURVES/SURFACES
• Advantages
• Easy to determine inside/outside
• Easy to determine if a point is on the surface

• Disadvantages
• Hard to generate points on the surface

• Dr. Scott Schaefer

8
EXPLICIT FORM
http://www.robots.ox.ac.uk/~ian/Teaching/CompGeom/lec3.pdf

9
 PRINCIPAL OF COMPUTER GRAPHICS
The points are used to form geometries.
Large increment of t ➔discontinuous look
➢Line ➔
x = 𝑥1 + 𝑥2 − 𝑥1 𝑡 0≤𝑡≤1
y = 𝑦1 + 𝑦2 − 𝑦1 𝑡
Symmetric digital differential analyzer (DDA)

10
PARAMETRIC CURVES/SURFACES
P(t) = (x(t), y(t), z(t))
P(s,t) = (x(s,t), y(s,t), z(s,t))

• Dr. Scott Schaefer

Geri’s Game Copyright Pixar

11
 CIRCLE
• Cartesian equation: x2 + y2 = r2

• Parametrical equation:
x = r cos(t), y = r sin(t)

Copyright @ 2002 by Jim X. Chen: jchen@cs.gmu.edu

12
PARAMETRIC SURFACES

x(s,t) = 1+s2s2 +t 2
y(s,t) = 1+s2t2 +t 2
z(s,t) = 1−s2 −t 2
1+s2 +t 2

• Dr. Scott Schaefer

13
 PARAMETRIC CURVES/SURFACES
• Advantages
• Easy to generate points on the curve/surface
• Separates x/y/z components

• Disadvantages
• Hard to determine inside/outside
• Hard to determine if a point is on the curve/surface

• Dr. Scott Schaefer

15
SYNTHETIC CURVES
1. Parametric Cubic Spline (Hermite Cubic Spline)
2. Bezier curves
3. B-spline curves
4. Rational B-splines
5. Nonuniform rational B-splines

17
 APPROXIMATION VS INTERPOLATION
• Instead of fitting exact curves to points, approximated
curve is found.
• Used in airplane fuselage, ship hull, car body where
smooth surfaces are needed.

18
PARAMETRIC CUBIC (HERMITE)
where 𝑃(𝑡) is a point on the curve.

• 𝑃(𝑡) = 𝑎 3 𝑡 3 + 𝑎 2 𝑡 2 + 𝑎1𝑡 + 𝑎0 (Hermite interpolation)

• 𝑎0, 𝑎1, 𝑎2,𝑎3 are unknowns.

• 𝑃 0 = 𝑎0
• 𝑃 1 = 𝑎3 + 𝑎2 + 𝑎1 + 𝑎0 𝑃(0)′ 𝑷(𝒕)
• 𝑃′ 0 = 𝑎1
𝑃(1)
• 𝑃′ 1 = 3𝑎3 + 2𝑎2 + 𝑎1 𝑃(0)
𝑃(1)′
Geometric
coefficients

19
 PARAMETRIC CUBIC
• Unknowns are found:
- 𝑎0 =𝑃 0
- 𝑎1 = 𝑃′ 0
- 𝑎2 = −3𝑃 0 + 3𝑃 1 − 2𝑃′(0) − 𝑃′(1)
- 𝑎3 = 2𝑃 0 − 2𝑃 1 + 𝑃 ′ 0 + 𝑃′(1)

• 𝑃 𝑡 = 2𝑡 3 − 3𝑡 2 + 1 𝑃 0 + −2𝑡3 + 3𝑡2 𝑃 1 + 𝑡 3 − 2𝑡 2 + 𝑡 𝑃′ 0
+ 𝑡 3 − 𝑡 2 𝑃′(1)

20
PARAMETRIC CUBIC

• Matrix form:

𝑎3
𝑎2
- 𝑃 𝑡 = 𝑡3 𝑡2 𝑡 1 𝑎1
𝑎0

2 −2 1 1 𝑃(0)
- 𝑃 𝑡 = 𝑡3 −3 3 −2 −1 𝑃(1)
𝑡2 𝑡 1 0 0 1 0 𝑃′(0)
1 0 0 0 𝑃′(1)

21
EXAMPLE
Find geometric coefficients of parametric cubic curve using following
information:
for 𝑡 = 0 ⇒ 2,20,2 is a point on the curve and 𝑃′ 0 = (𝑥1 , 0, 4𝑥1 )
for 𝑡 = 1 𝑖ç𝑖𝑛 ⇒ 10,20,2 is a point on the curve and 𝑃′ 1 = (𝑥2 , 0, −2𝑥2 )
for 𝑡 = 0.5𝑖ç𝑖𝑛 ⇒ 6, 20, 6 is a point on the curve

22
CONT.

2 −2 1 1 2 20 2
−3 3 −2 −1 10 20 2
6 20 6 = 0.53 0.52 0.5 1 𝑥1 0 4𝑥1
0 0 1 0
1 0 0 0 𝑥2 0 −2𝑥2

6 = 6 + 0.125𝑥1 − 0.125𝑥2
6 = 2 + 0.5𝑥1 + 0.25𝑥2
16
𝑥1 = 𝑥2 =
3

2 20 2
Geometric Coefficients 10 20 2
16/3 0 64/3
16/3 0 −32/3
23
 CONTINUITIES
• Smoothness of the connection of two curves or surfaces
at the connection points or edges.
• 𝐶0 continuity: simply connecting two curves.
• 𝐶1 continuity: gradients (tangent) at the point of joining
are same.
• 𝐶2 continuity: curvature continuity at the point of joining
is same.

24
BEZIER CURVES
• Employing control points (ordered set of points (𝑉0, … ,𝑉𝑛)).
• A Bezier curve of degree 𝑛, specified by 𝑛 + 1 control
vertices.
• Parametric form is as follows:

* the vectors 𝑉𝑖 represents the 𝑛 + 1 control points.

* the blending function
* 𝐵𝑖,𝑛 𝑡 ≥ 0 for all 𝑖.
𝑛

෍ 𝐵𝑖,𝑛 𝑡 = 1 0≤𝑡≤1
𝑖=0
25
BEZIER CURVES
• Previous page conditions force the curve to lie entirely within
the convex hull of the curve.

• 𝑄 𝑡 = 𝑉0𝐵0,3 + 𝑉1𝐵1,3 + 𝑉2𝐵2,3 + 𝑉3𝐵3,3

• 1 − 𝑡 3 + 3𝑡 1 − 𝑡 2 + 3𝑡2 1 − 𝑡 + 𝑡 3 = 1
• 𝑄 𝑡 = 1 − 𝑡 3𝑉0 + 3𝑡 1 − 𝑡 2𝑉1 + 3𝑡2 1 − 𝑡 𝑉2 + 𝑡3𝑉3

• 𝑡 = 0 → 𝑄 0 = 𝑉0.
• 𝑡 = 1 → 𝑄 1 = 𝑉3. 𝑄′ 0 = ?
26
BEZIER CURVES
• Plot of blending functions for a cubic Bezier curve.

𝐵3,3
𝐵0,3

𝐵1,3 𝐵2,3

• At 𝑡 = 0, 𝐵0,3 = 1, other blending functions are 0.

• At 𝑡 = 1, 𝐵3,3 = 1, other blending functions are 0.
27
BEZIER CURVES
• Matrix form:

−1 3 −3 1 𝑉0
3 −6 3 0 𝑉1
- 𝑄 𝑡 = 𝑡3 𝑡2 𝑡 1
−3 3 0 0 𝑉2
1 0 0 0 𝑉3

28
 EXAMPLE
Bezier curve

Find the points on the Bezier curve for t =

0, ¼, ½ , ¾ and 1. The coordinates of
the four control points are given by:

29
 BEZIER CURVES
• Closed curves can be obtained by making the first and
last control points coincide.

• Bezier curves lack local control.

30
BEZIER CURVES
• 𝐶1 continuity is usually applied to piecewise Bezier
curves.

• The connecting control point 𝑉4 (which is also 𝑉0∗ ) joins 𝑉3

and 𝑉1∗ in a straight line.

31
BEZIER CURVES
• Changes on a Bezier curve caused by
repositioning of control points.

32
Example

𝑄1′ 1 = 𝑄2′ 0

3 3−𝑥 4−𝑦 3−𝑧 = 3[ 2 − 3 6 − 4 (0 − 3)] 𝑥=4 𝑦=2 𝑧=6 33

CONT.

34
B-SPLINE CURVES
• Exhibiting local control.
• The curvature continuity 𝐶 2 is guaranteed by cubic B- splines.

Recall that lack local control in Bezier!

35
GENERAL B-SPLINE REPRESENTATION
• General form:

where 𝑉𝑖 is set of control points and 𝑁𝑖,𝑘(𝑡) is the blending function of

degree (𝑘 − 1).
• Order is 𝑘, degree (𝑘 − 1) and 𝐶 𝑘−2 continuous.

• For example;
-𝑘= 4
- Degree = 𝑘 − 1 = 3
- Second degree continuity is satisfied.

36
 GENERAL B-SPLINE REPRESENTATION
• The i th blending function 𝑁𝑖,𝑘(𝑡) is:

where the knot vector is [𝑡𝑖, ⋯ ,𝑡𝑖+𝑘].

• The constraints imposed on the knot vector are:
- Nondecreasing order 𝑡𝑖 ≤ 𝑡𝑖+1.
- The same value should not appear more than 𝑘 times.
• Normalizing property:

• Degree (𝑘 − 1), 𝑛 + 1 control points, a knot vector [𝑡0, ⋯ ,𝑡 𝑚 ].

- 𝑚 +1 = 𝑛+1 +𝑘

37
UNIFORM CUBIC B-SPLINE CURVES
• The uniform cubic B-spline blending function 𝑁𝑖(𝑡) is a cubic 𝐶 2 basis function.
• The parametric intervals (or knots) 𝑡, within which the basis function is defined,
are equal.

[𝑡𝑖, 𝑡𝑖+4].

38
 UNIFORM CUBIC B-SPLINE CURVES
• Uniform cubic B-spline curve with six control points:
- 𝑉0, 𝑉1, 𝑉2, 𝑉3, 𝑉4, 𝑉5
• Three curve segments:
- Seg1 → 𝑉0, 𝑉1, 𝑉2,𝑉3
- Seg2 → 𝑉1, 𝑉2, 𝑉3,𝑉4
- Seg3 → 𝑉2, 𝑉3, 𝑉4,𝑉5 Seg1
Seg2
Seg3

• 𝑁0,3 𝑡 + 𝑁1,3 𝑡 + 𝑁2,3 𝑡 + 𝑁3,3 𝑡 = 1 0 ≤ 𝑡 ≤ 1. 40

UNIFORM CUBIC B-SPLINE CURVES
• To determine the basis function, four segments by a general cubic
polynomial of the form is defined:
- 𝑁𝑗,3 𝑡 = 𝑎𝑗 + 𝑏𝑗𝑡 + 𝑐𝑗𝑡2 + 𝑑 𝑗𝑡3 where 𝑗 = 0,1,2,3.
• 16 unknown polynomial coefficients should be found for each of the
segment.

• 𝑁0,3 0 = 𝑁 ′ 0,3 0 = 𝑁 ′′ 0,3 0 = 0 at the left end.

• 𝑁3,3 1 = 𝑁 ′ 3,3 1 = 𝑁 ′′ 3,3 1 = 0 at the right end.

• 𝐶 0 , 𝐶 1 , 𝐶 2 continuity requirements at the three internal knots

• 𝑁0,3 𝑡 + 𝑁1,3 𝑡 + 𝑁2,3 𝑡 + 𝑁3,3 𝑡 = 1 where 0 ≤ 𝑡 ≤ 1.

41
UNIFORM CUBIC B-SPLINE CURVES

• Solving these sixteen equations.

1
• 𝑁0,3 𝑡 = 𝑡 3 0≤𝑡≤ 1
6

• 𝑁1,3 𝑡 = 1 (−3𝑡3 + 3𝑡2 + 3𝑡 + 1)

6

• 𝑁2,3 𝑡 = 1 (3𝑡3 − 6𝑡2 + 4)

6

• 𝑁3,3 𝑡 = 1 (−𝑡3 + 3𝑡2 − 3𝑡 + 1)

6

42
UNIFORM CUBIC B-SPLINE CURVES

• Matrix form:

−1 3 −3 1 𝑉𝑖 −1
1
1 3 −6 3 0 𝑉𝑖
-𝑃 𝑡 = 𝑡3 𝑡2 𝑡
6 −3 0 3 0 𝑉𝑖 +1
1 4 1 0 𝑉𝑖 +2

43
UNIFORM QUADRATIC B-SPLINE CURVES
• Matrix form:

1 −2 1 𝑉𝑖 −1
1
- 𝑃 𝑡 = 𝑡2 𝑡 1 −2 2 0 𝑉𝑖
2
1 1 0 𝑉𝑖 +1

44
UNIFORM QUADRATIC B-SPLINE CURVE
• Closed quadratic B-spline curve produced by six control
points with six segments.

45
 EXAMPLE
Find starting and ending
locations of uniform quadratic
B-spline curve

46
EXAMPLE

47
CONVERSION BETWEEN REPRESENTATIONS
• [𝑀] 𝑓𝑟𝑜𝑚 [𝑉] 𝑓𝑟𝑜𝑚 = [𝑀]𝑡𝑜 [𝑉]𝑡𝑜

• [𝑉] 𝑡𝑜 = 𝑀 −1 [𝑀]𝑓𝑟𝑜𝑚 [𝑉]𝑓𝑟𝑜𝑚

𝑡𝑜

From Bezier to B-spline

𝑀 −1
𝑡𝑜 [𝑀] 𝑓𝑟𝑜𝑚

48
EXAMPLE

49
 CONT.
(a)

(b) −6 0 0 1
6 0 0 1
𝑉 𝐻 =
9 12 0 0
9 12 0 0
50
UNIFORM/PERIODIC B-SPLINES
• Equispaced knot values; [ 𝑡𝑖 − 𝑡𝑖−1 = 𝑎].
• For example;
- 0 1 2 3 4 with 𝑎 = 1.
- [−0.5 0.0 0.5 1.0 1.5] with 𝑎 = 0.5.

Periodic functions for a uniform B-spline

(The functions are a shifted version of the
previous one).

51
 NONUNIFORM B-SPLINES
• Multiplicity of knot values occurred at the ends and the internal knots
were equispaced.
• For example;
- [0 1 2 3 3 4]
- [0.0 0.20 0.55 0.75 1.0]

52
 NONPERIODIC B-SPLINES

• Nonperiodic B-splines of various degrees.

• Cubic nonperiodic B-spline exhibiting local control.

53
EFFECT OF MULTIPLICITY OF CONTROL POINTS

54
NON UNIFORM RATIONAL B-SPLINES (NURBS)
• Rational curves with varying weight at 𝑉4.

where
- 𝐵𝑖𝑛(𝑡)=standard Bezier blending function
- 𝑤𝑖=weights at each control point.

Bezier B-Spline
𝑛 𝑛

𝑄 𝑡 = ෍ 𝐵𝑖𝑛 𝑡 𝑉𝑖 𝑃 𝑡 = ෍ 𝑁𝑖𝑘 𝑡 𝑉𝑖
Nonrational 𝑖=0
𝑖=0

σ𝑛𝑖=0 𝐵𝑖𝑛 𝑡 𝑤𝑖 𝑉𝑖 σ𝑛𝑖=0 𝑁𝑖𝑘 𝑡 𝑤𝑖 𝑉𝑖

Rational 𝑄 𝑡 = 𝑛
σ𝑖=0 𝐵𝑖𝑛 𝑡 𝑤𝑖
𝑃 𝑡 = 𝑛
σ𝑖=0 𝑁𝑖𝑘 𝑡 𝑤𝑖

55
NON UNIFORM RATIONAL B-SPLINES (NURBS)

57
SUMMARY
• (i-1) for t=0
• (i-1) for t=1
• (i) for t=0
𝑃𝑖−1 (𝑖 − 1) segment
𝑃′𝑖−1
• (i) for t=1
𝑃𝑖 (𝑖) segment
𝑃′𝑖
𝑃𝑖+1
𝑃′𝑖+1

𝑃𝑖−1 (𝑖 − 1) segman
𝑃′𝑖−1
𝑃′′𝑖 −1
𝑃𝑖 (𝑖) segman
𝑃′𝑖
𝑃′′𝑖 𝑃𝑖+1
𝑃′𝑖+1
Cubic b-splines 𝑃′′𝑖 +1
58
 SUMMARY Q,V,P→ all points Geometric coefficients

Parametric cubic curve

(Hermite interpolation)

Cubic Bezier curve

Cubic B-spline curve

Quadratic B-spline curve

59
SURFACES
60
SURFACE MODELING

61
USE OF SURFACE MODELS
•Checking for interference
•Between mating parts
•Generating cross-section views
•Generating finite elements meshes
•Generating NC tool paths for machining
•Etc.

62
 IMPLICIT AND EXPLICIT SURFACES
• There are two types of surfaces that are commonly used in
modeling systems, parametric and implicit.

• Implicit surface: 𝑓(𝑥, 𝑦, 𝑧) = 0

(𝑥 − 𝑥0)2 + (𝑦 − 𝑦0)2 + (𝑧 − 𝑧0)2 − 𝑟2 = 0

• Parametric surfaces are defined by a set of three functions,

one for each coordinate:

𝑥 = 𝑓(𝑠, 𝑡), 𝑦 = 𝑓(𝑠, 𝑡), 𝑧 = 𝑓(𝑠, 𝑡)

63
 CURVES VS SURFACES

• Parametric curve equation:

- 𝑃(𝑡)= [𝑥(𝑡) , 𝑦(𝑡),𝑧(𝑡)].
• Parametric surface equation:
- 𝑃 𝑠,𝑡 = [𝑥 𝑠, 𝑡 , 𝑦 𝑠, 𝑡 , 𝑧(𝑠,𝑡)].

64
 SURFACES OF REVOLUTION
➢Rotating a plane curve around an axis.

65
 SURFACES OF REVOLUTION R(θ)

➢Any point on the surface of revolution is a function of two

parameters, 𝑡 and 𝜃.
 Parameter 𝑡 describes the entity to be rotated.
 𝜃 is the angle of rotation.

➢A point on the line segment is represented by L(t)= [𝑥(𝑡), 0, 𝑧(𝑡)].

➢When that point is rotated by 𝜃, it becomes:
P(t, θ) = L(t) R(θ)
P(t, θ)=[𝑥(𝑡) cos 𝜃, 𝑥(𝑡) sin 𝜃 , 𝑧(𝑡)]
➢ Matrix form:
cos𝜃 sin𝜃 0 0
- P 𝑡, 𝜃 = 𝑥(𝑡) 0 𝑧(𝑡) 1 0 0 0 0
0 0 1 0
0 0 0 1
66
EXAMPLE
Generate the conical surface obtained by rotation of the line segment AB with
𝐴 = (1,0,1) and 𝐵 = (7,0,7) around the z axis.

AB Line Equation= A + (B-A)t

67
SURFACES OF REVOLUTION
➢ A simple torus can be generated by rotating a circle contained
in the xz plane around the z axis.
➢ Parametric form of circle: [𝑎 + 𝑟 cos 𝜙 , 0, 𝑎 sin𝜙] .
➢ Parametric form of a torus:

68
 EXAMPLE

result
Cubic Bezier Curve
Requesed Point's Revolution
parameters
69
SWEEPING
➢Surface is generated through the movement of a
line/curve/polygon/etc. along or around a path.

70
SWEEPING
• The swept surface is represented in parametric form as:
- 𝑃 𝑡,s = 𝑄 𝑡 [𝑇(𝑠)]
* 𝑄(𝑡) is the parametric equation of the line or curve.
* [𝑇(𝑠)] is the sweep transformation based on the path.

• If the path is a straight line starting at the origin, 𝑥 𝑠 = 𝑎𝑠,

𝑦 𝑠 = 𝑏𝑠, 𝑧 𝑠 = 𝑐𝑠.
1 0 0 0
- 𝑇 𝑠 = 0 1 0 0 0 ≤ 𝑠 ≤ 1.
0 0 1 0
𝑎𝑠 𝑏𝑠 𝑐𝑠 1

• If the path involves rotations about the coordinate axes,

then the appropriate rotation angles should be included in
the sweep transformation.
71
 EXAMPLE
Generate the swept surface obtained by moving a
triangle located in the xy plane along the path 𝑧 = 8𝑠,
𝑦 = sin 𝜋𝑠 − 1 . The vertices of the triangle are
𝑃1 = (−5,0,0), 𝑃2 = (0,5,0), 𝑃3 = (5,0,0). Keep the
normal to the triangle in the direction of the
instantaneous tangent to the path.

72
CONT. 1 0 0 0
−5 0 0 1
0 0.964 0.268 0
𝑃 0.25,15.52° = 0 5 0 1
0 −0.268 0.964 0
5 0 0 1
0 −0.293 2.0 1

−5.0 −0.293 2.0 1

= 0 4.525 3.338 1
5.0 −0.293 2.0 1

73
FREEFORM SURFACES
➢Some surfaces cannot be totally described by
the analytical representations.
➢For example: automobile bodies, ship hulls,
etc…
➢Series of “patches” are used to represent
models.

74
PARAMETRIC CUBIC SURFACES
➢ Hermite curve formulation:
2 −2 1 1 𝑃(0)
1] −3 3 −2 1 𝑃(1)
- 𝑃 𝑡 = [𝑡3 𝑡2 𝑡
0 0 1 0 𝑃′(0)
1 0 0 0 𝑃′(1)
-𝑃 𝑡 = 𝑡 𝑀 𝐻 𝐺 𝐻

➢ Similar equation is used in both direction s and t directions:

- P(s, t) = [𝑠][𝑀]𝐻[𝐺]𝐻[𝑀]𝐻𝑇 [𝑡]𝑇

75
 PARAMETRIC CUBIC SURFACES
• [𝐺]𝐻 matrix represents the following:
[𝐺]𝐻 parameters
𝑃(0,0) 𝑃(0,1) 𝑃𝑡 (0,0) 𝑃𝑡(0,1)
𝑃(1,0) 𝑃(1,1) 𝑃𝑡 (1,0) 𝑃𝑡(1,1)
- [𝐺]𝐻 =
𝑃𝑠 (0,0) 𝑃𝑠(0,1) 𝑃𝑠𝑡(0,0) 𝑃𝑠𝑡(0,1)
𝑃𝑠 (1,0) 𝑃𝑠(1,1) 𝑃𝑠𝑡(1,0) 𝑃𝑠𝑡(1,1)

𝐷𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑤𝑖𝑡ℎ
𝑃𝑜𝑠𝑖𝑡𝑖𝑜𝑛 𝑜𝑓
𝑟𝑒𝑠𝑝𝑒𝑐𝑡 𝑡𝑜 𝑡 𝑎𝑡
𝑐𝑜𝑟𝑛𝑒𝑟 𝑝𝑜𝑖𝑛𝑡𝑠
𝑐𝑜𝑟𝑛𝑒𝑟 𝑝𝑜𝑖𝑛𝑡𝑠
- [𝐺]𝐻 = 𝐷𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑤𝑖𝑡ℎ
𝐶𝑟𝑜𝑠𝑠 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
𝑟𝑒𝑠𝑝𝑒𝑐𝑡 𝑡𝑜 𝑠 𝑎𝑡
𝑎𝑡 𝑐𝑜𝑟𝑛𝑒𝑟 𝑝𝑜𝑖𝑛𝑡𝑠
𝑐𝑜𝑟𝑛𝑒𝑟 𝑝𝑜𝑖𝑛𝑡𝑠

76
PARAMETRIC CUBIC SURFACES
Variation of twist vectors Variation of tangent vectors
changing the shape. changing the shape.

77
PARAMETRIC CUBIC SURFACES
Normal vector at a point on the surface is computed as follows:
- 𝑛 = 𝑃′𝑠 × 𝑃′𝑡

78
 EXAMPLE

Cartesian components are

given for a parametric cubic
surface. Find tangent
vectors for t = 1 and s = ½.

79
CONT. t = 1 ve s = ½

80
BEZIER SURFACES

81
 BEZIER SURFACES
➢The surface takes the general shape of the control points.
➢The surface is contained within the convex hull of the CPs.
➢The surface corners and corner CPs are coincident.

82
BEZIER SURFACES
➢𝐶1 continuity are enforced along the common edge between two
patches.
➢Force the polygon control sides meeting at the common edge to
be collinear.

83
B-SPLINE SURFACES

• Parametric equation:

* 𝑉𝑖,𝑗 are the control points.

* 𝑁𝑖,𝑘(𝑠), 𝑁𝑗,𝑙 (𝑡)are the B-spline blending functions.

84
B-SPLINE SURFACES

• Bicubic surface are represented by a 4x4 array of control

points.

Local control

85
B-SPLINE SURFACES
For 𝐶2 continuity, 16 control points are needed for each patch,
twelve are shared with the next segment in developing the surface.

86
B-SPLINE SURFACES
➢Cylindrical surface ➢Closed curve

Translate the control points of the

closed curve (circle) in a direction
perpendicular to the plane.

87
B-SPLINE SURFACES

➢Torus formed by B-spline

patches.
▪ Rotation around an
axis.

88
EXAMPLE

89
EXAMPLE

90
CONVERSION BETWEEN CUBIC SURFACE REPRESENTATIONS

• 𝑉𝑡𝑜 = [𝑀𝑡𝑜−1𝑀𝑓𝑟𝑜𝑚]𝑉𝑓𝑟𝑜𝑚[𝑀𝑡𝑜−1𝑀𝑓𝑟𝑜𝑚]𝑇

91
 RATIONAL SURFACES
• 𝑉𝑖,𝑗 are the control points.
• 𝑁𝑖,𝑘(𝑠), 𝑁𝑗,𝑙 (𝑡)are the B−spline blending
functions.
• 𝑤𝑖,𝑗 are the weights at each control point.

Bezier B-Spline
𝑛 𝑚 𝑛 𝑚

Non-rational 𝑄 𝑠, 𝑡 = ෍ ෍ 𝐵𝑖,𝑛 𝑠 𝐵𝑗,𝑚 (𝑡)𝑉𝑖,𝑗 𝑃 𝑠, 𝑡 = ෍ ෍ 𝑁𝑖,𝑘 (𝑠) 𝑁𝑗,𝑙 𝑡 𝑉𝑖,𝑗

𝑖=0 𝑗=0 𝑖=0 𝑗=0

σ𝑛𝑖=0 σ𝑚
𝑗=0 𝑤𝑖,𝑗 𝐵𝑖,𝑛 𝑠 𝐵𝑗,𝑚 (𝑡)𝑉𝑖,𝑗 σ𝑛𝑖=0 σ𝑚
𝑗=0 𝑤𝑖,𝑗 𝑁𝑖,𝑘 (𝑠) 𝑁𝑗,𝑙 𝑡 𝑉𝑖,𝑗
Rational 𝑄 𝑠, 𝑡 = 𝑛 𝑃 𝑠, 𝑡 =
σ𝑛𝑖=0 σ𝑚
σ𝑖=0 σ𝑚𝑗=0 𝑤𝑖,𝑗 𝐵𝑖,𝑛 𝑠 𝐵𝑗,𝑚 (𝑡) 𝑗=0 𝑤𝑖,𝑗 𝑁𝑖,𝑘 (𝑠) 𝑁𝑗,𝑙 𝑡
92
NURBS
Non-uniform Rational B-spline Surfaces

93
Additional Surface Entities
94
PLANE SURFACE
• Three noncoincident points can define a plane surface.

P(u,v) = P0 + u(P1 − P0) + v(P2 − P0)

0  u1 0  v 1

95
RULED SURFACE
• Linear interpolation between two curves.

v u constant
0

v G ( U ) − P( u , v)
=
1 G(U ) − Q(U )
P(u, v) = (1− v)G(u) + vQ(u)
0  u 1 0  v 1
96
 TABULATED CYLINDER
• Sweeping a planar curve in space in a certain direction at a certain distance.

v is the cylinder length

n is the cylinder axis

0  u umax
P(u, v) = G(u) + vn
0  v vmax
97
 REFERENCES
This presention is a compilation of following sources:

1. Assoc. Prof. Dr. Erkan Gunpinar, CAD/CAM Lecture Notes,

Istanbul Technical University.
2. Chang, T.C. Wysk, R. A. and Wang, H. P., Computer-Aided
Manufacturing. Prentice Hall, 2nd Ed., 1997

98