Ntedi These ant reGire ol becau thy aUe

coMen vetos
to be reprsnted
COMPUTER GRAPHICS AND FUNDAMENTALS OF IMAGE PROCESSING
MODULE 2

Homogeneous Coordinates:
Multiplicative and translational terms for a
can be combined into asingle matrix if we two-dimensional geometric transformation
We can use the third column of a expand the representations to 3 x 3 matrices
all transformation equations can betransformation matrix for the translation terms, and
We also need to expand the matrixexpressed as matrix multiplications.
position to a three-element column matrixrepresentation for a two-dimensional cOordinate
A standard technique for
accomplishing this is to expand each two dimensional
coordinate-position representation
called homogeneous coordinates, where
(x, y) toa three-element
representation (xh, yh, h),
nonzero value such that
the homogeneous parameter h is a

y=
A general
two-dimensional homogeneous coordinate representation could also be
written as (h-x, hy, h).
A convenient choice is simply to set h
= 1. Each
represented with two-dimensional position is then
homogeneous coordinates (x,y, 1).
The term homogeneous coordinates is used in mathematics to refer to the
this representation on Cartesian effect of
equations.
Two-Dimensional Translation Matrix:
The
homogeneous-coordinate for translation is given by
0I
" This translation operation can be written in the
abbreviated form
P=T(r,ty) "P
with T(tx,ty) as the 3 x 3 translation matrix
Two-Dimensional Rotation Matrix:
Two dimensional rotation transformationequations about the coordinate origin can be
expressed in the matrix form
cOs -sint
sin

P= RØ) ·P
The rotation transformation operator R(8 ) is the 3 x 3
matrix with rotation parameter
.
projeotin
COMPUTER GRAPHICS AND FUNDAMENTALS OF AGE PROCESSING NOD01E 2

Two-Dimensional Scaling Matrix:

A scalingtransformation relative to the coordinate origin can now be expressed as the
matrix multiplication

P'=S(s,, S,)· P J3rd

The scalingoperator S(sx, sy) is the 3 x 3 matrix with parameters sx and sy
General Two-Dimensional Fixed-Point Scaling:
respect to a selected fixed position (x1, y).
" To produce atwo-dimensionalscaling with relative
when we have a function that can scale to the COordinate origin only. This
sequence is

(d}
(a) Transtate Object
Original Position Translat Object Scale Object so that the Fxed
of Objectand so that Fixed Point with Respect Pont is Retuned
Fixed Point (p y) is at Origin ta Crigin
to Position ( y

the coordinate origin.

1. Translatethe objectso that the fixed point coincides with
2. Scale the object with respect to the coordinate origin.
to its original
3. Use the inverse of the translation in step (1) to return the object
position.

required scaling
Concatenating the matrices for thesethree operations produces the
matrix

[1 0 x] fs, 0 0] 1 0 -x] 0 x, (l - s)]

1
L0 0 1

T(af, yf) "S(8z, Sy)- T(-xf,-y) =S(xf, Vf, Sx. Sy)

CONIPRUTER GRAPHISAND FUNDAMENTALSOF MAG: EROcESSNG
NODULE2
General Two-Dimensional Scaling
Directions:
Parameters sx and sy scale objects along the Xand y
We can scale an object in other directions.
directions by rotating the object to align the
Scaling directions with the coordinate axes before desired
transformation. applying the scaling
Suppose we want to apply scaling factors with values
s2 in the directions shown in Figure specified by parameters s1 and

" The composite matrix resulting from the product of these three transformations is
TSi cos + S) sin' e (S -`)cos e sin @01
R-(e)-S($1, S2) ·-R(@)
(S1, $2) R(e) =
=($- S) cos @sin 0 Ssine+$2 cos² e 0
0 0
Matrix Concatenation Pronerties:
mensional Transformations:
The matrix representation for this reflection relative to the xy plane is
Composite Three-Dimensional Transformations 1 0
" We form a composite three dimensional transformation by multiplying the matrix 01
Mereilkct =
representations for the individual operations in the transformation sequence. 00 -1 0
We can implement a transformation sequence by concatenatíng the individual
matrices from right to left or from left to right, depending on the order in which the
matix representations are specified
Three-Dimensional Shears
’ These transformations can be used to
-Dimensional Reflections: modify object shapes.
’ For three-dimensional we can also
Areflection in a three-dimensional space can be generate shears relative to the z axis.
performed relative to a selected ’A general z-axis shearing transformation relative to a sele
reflection axis or with respect to a reflection plane. cted reference
Reflections with respect to a plane are similar; when the reflection plane is a position is produced with the following matrix:
plane (y, Xz, Or yz), we can think of the transformation as a 180" coordinate
rotation in four
dimensional space with a Conversionbetween a left-handed frame and a
frame right-handed 0 shzr -shzr Zet
An example of a reflection that converts coordinate
specifications froma right handed Meshear = 00 01 shay
1
-shey Zref
system to a left-handed system is shown below
0 0 0 1
Reflection
Relative to the ’ The Below figure shows the shear
transformation of a cube
xy Plane

Aunit cube (a) is sheared relative to the oriqin (b) by Matrix 46, with shzx
Three-Dimensional Translation

A
position P=(x. y, z) in three-dimensional space is translated to a location P= (x
Vz)by addingtranslation distances tx, ty, and tz to the Cartesian coordinates of P:
x=!+fr

We can express these three-dimensional translation operations in matrix form

Or

P=T.P

ty, tz )
Moving acoordinate position with translation vectorT=(tx,
y axis

e(x',y'.z')

T=(y )

the positionof a three-dimensional object using translaion vector T

Shifting
 Three-Dimensional Scaling:
Thematrix expression for the three-dimensional scaling transformation of a position P
= (x, y. z) is given by

The three-dimensional scaling transformation for a point position can be represented

as

P=S P
where scaling parameters sX, sy, andsz are
assigned anypositive values.
ATAANDA EHAI
 OFMAGE PROCEsSTNG VODULE2
RGRAPHISAND EUNDANENTALS

relative to the origin are

Explicit expressions for the scaling transformation
x'=!: Sx, V'=ySy
routine that scales relative to
Because some graphics packages provide only a
scaling transformation with respect
the coordinate origin, we can always construct a transformation sequence:
toany selected fixedposition (xf,yf , zf )using the following
1. Translate the fixed point to the origin.
coordinate origin
2 Apply the scaling transformation relative to the
position.
Translate the fixed point back to its original
demonstrated
This sequence of transformations is

0 (1-Sy)xf
Sy (1- Sy)yf
T(xf, yf, zf)- S(S,, Sy, S,)· T(-xf,-yf,-2f)= 0 S, (1- S,)2f
1

invese Irasat