You are on page 1of 3

CURVES

The most preferred form of representation of curves adopted in modeling is the parametric form,

x=x(t), y=y(t), z=z(t)

where (x, y, z) is a point on the curve corresponding to a unique value of the parametric variable t.
The function x(t) stands for some as yet unspecified polynomial function of degree n in t as
2 3 n
x (t)=ax 0 +a x 1 t +a x 2 t + a x3 t +.........+a xn t . similarly for y(t) and z(t). where n can be any non-
negative integer, the nature and shape of curves varies significantly with the value of n. We get
straight line, when n=1; a quadratic plane curve when n=2; a cubic space curve when n=3.

We will study the mostly used cubic polynomial curve segment as,

P(t)=[ x (t) y ( t) z (t)]


3 2
x (t)=ax t +b x t +c x t + d x
y (t)=a y t 3 +b y t 2 +c y t+d y
3 2
z (t)=az t +b z t +c z t +d z , 0⩽t⩽1

To deal with finite segment of the curve, without loss of generality, we restrict the parameter t to the
[0, 1] interval. The values for the coefficient t in the parametric equations are determined from the
boundary conditions on the curve sections. One boundary condition is that two adjacent curve
sections have the same coordinate position at the boundary, and a second condition to match the two
curve slopes at the boundary so that we obtain one continuous, smooth curve. Continuous curves
that are formed with polynomial pieces are called spline curve or spline.

We can combine x(t), y(t) and z(t) and their respective constant coefficients (a, b & c terms) into an
equivalent, more concise, vector equation P(t)=⃗a t 3 + b⃗ t 2 + c⃗ t+ d⃗ , 0⩽t⩽1

The constant coefficient are in vector form (not a geometric points)


a⃗ =[a x a y a z ]

b=[b x b y bz ]
c⃗ =[c x c y c z ]
⃗d=[d x d y d z ]

SPLINE CURVE
The term spline refers to any composite curve formed with polynomial sections satisfying specified
continuity conditions at the boundary of the sections.

Depending on the type of polynomial used and set of boundary conditions satisfied, the following
varieties of spline curves are used in graphic application.
1. Piecewise Cubic Spline: HERMITE SPLINE
2. BEZIER SPLINE
3. B-SPLINE

Interpolated Spline: When polynomial sections are fitted so that the curve passes through each
control point, the resulting curve is said to have interpolated the set of control points.

Approximated Spline: When polynomial sections are so fitted that without necessarily passing any
of the control points, it approximates the shape of the control polygon, the resulting curve is said to
have interpolated the set of control points.

HERMITE CURVE

It is a piecewise interpolated cubic spline.


Let P(t)=⃗a t 3 + b⃗ t 2 + c⃗ t+ d⃗ … … (1)
represents any point P(t)=[ x (t) y ( t) z (t)] on a Hermite Curve section. In matrix formulation
this curve piece can be expressed as

[]
a⃗
⃗b
P (t)=[t 3 t 2
⃗ t 1] , 0⩽t⩽1
c⃗
⃗d
= [T ][ A]
The four unknown algebraic coefficient ⃗a , ⃗ b , ⃗c , ⃗d are determined by four boundary
conditions including two end points P1 (t=0) and P2 (t=1) and two parametric target vectors P11 and
P21 at end points P1 and P2 respectively.
The parametric first derivative P1(t) is the parametric tangent vector at P(t) on the curve applying
this definition to the curves equation, we have
d d d d
P (t)=P' (t)=[ x (t ) y (t ) z (t)]
dt dt dt dt
d
= [T ] [ A]
dt
= [3 t 2 2 t 1 0 ][ A ]
= 3 ⃗a t 2 +2 ⃗b t+ ⃗c
Evaluating P(t) and P1(t) at t = 0 and t = 1 we get,
P(t=0)=P1= ⃗d
P(t=1)=P2=⃗a + b+ ⃗ ⃗c + ⃗d
P1 (t=0)=P'1= ⃗c
P(t=1)=P'2=3 a⃗ + 2 b⃗ + ⃗c

Solving the above four equations in four unknowns we get,


⃗a =2 ⃗ P 2+ P '1+ P'2
P1 − ⃗
⃗b=−3 ⃗ P2−2 P'1−P'2
P1 +3 ⃗
⃗c =P'1
d⃗ =P1
Substitute these results in equation (1)
P(t)=(2 P1−2 P2+ P '1+ P'2) t 3 +(−3 P1 +3 P2−2 P '1−P'2) t 2+ P'1 t+ P 1
Rearranging the terms we can rewrite the above expressions as,
P(t)=(2 t 3 −3 t 2 +1) P1 +(−2 t 3+3 t 2) P2 +(t 3−2 t 2+t ) P'1 +(t 3−t 2 )P'2
Again this can be written as,
P(t)=F 1 (t) P1 + F2 (t ) P2 + F 3 (t) P '1+ F 4 (t) P'2
= [F ][ B]
Where F1 (t ), F2 (t ), F3 ( t), F 4 (t) are referred to as Hermite blending function and are
defined by
F1 (t )=2 t 3−3t 2 +1
3 2
F2 (t )=−2t +3 t
3 2
F3 (t)=t −2 t +t
F 4 (t)=t 3−t 2
These functions of t are so named because they “blend” the contributions of the boundary constraint
values (P1, P2, P11 and P21) to create each point on the Hermite curve. Thus we see, any point on the
curve is a blended or weighted sum of the coordinate positions and tangent vectors at the boundary
of the curve.
We define blending function matrix [F ]=[ F1 (t ) F2 (t ) F 3 (t) F 4 (t )] and boundary

[]
P1
P
condition matrix [B]= 2'
P1
P'2
Now, [F ]=[ T ][M ]

[ ]
2 −2 1 1
−3 3 −2 −1
where, [T ]=[t 3 t 2 t 1] and [M ]= is the Hermite basis matrix, which
0 0 1 0
1 0 0 0
is fixed for any normalized Hermite spline from t = 0 to t = 1.

You might also like