Lecture # 04:

Imaging Geometry:
Imaging Geometry:

 Imaging Geometry plays a very important part in Image

processing, as it deals with the techniques and transformation
for mapping Real world objects and their characteristics on
Image plane.

 Any point in 3D co ordinate system is represented by BLOCK

letters, termed as World Coordinate system.
 W(X,Y,Z)

 Any point in 2D co ordinate system / Image plane is

represented by small letters, termed as Camera coordinate
system.
 f (x, y)
Translation:
Y
P2(X2,Y2,Z2)

(dx,dy,dz)

P1(X1,Y1,Z1)

Z
Translation:

 Suppose that the task is to translate a point P1, with

coordinate values (x1 , y1 , z1 ) to new location P2,by Using
displacement ( dx , dy , dz).

 The translation is easily accomplished by using the equation:

X2 = X1 + dx ===>(i)
Y2 = Y1 + dy ===>(ii)
Z2 = Z1 + dz ===>(iii)
Translation:

 Where (X2, Y2, Z2) are new points and can be represented in
Matrix form, as:

 X2 1 0 0 dx x1
y2 0 1 0 dy y1
z2 0 0 1 dz z1

V* = T V
where, V* is the new point,
T is the translation matrix,
and V is the original point.
Translation:
 Here, the product of Matrices T & V seems Invalid,
As Columns of T = Rows of V

 So we can generalize this transformation by inserting a DUMMY

row in all three matrices.
V* = T V
X2 100 dx x1
y2 010 dy y1
z2 00 1 dz z1
1 00 0 1 1

1 0 0 dx
0 1 0 dy
T = 0 0 1 dz Translation Matrix
0 0 0 1
Inverse Translation:

 The Initial Point P1 can be traced back by applying the Inverse

Translation, that is:

V = T-1
V* X1 1 0 0 -dx x2
y1 0 1 0 -dy y2
z1 0 0 1 -dz z2
1 000 1 1

1 0 0 -dx
T-1 = 0 1 0 -dy Inverse Translation
0 0 1 -dz Matrix
0 0 0 1

T . T-1 = I
Translation:
Original Image Translated Image
Translation:
Translated Image Translated Image

This property is used in adjusting the Objects on Monitor screen.

Scaling:

 The scale operator performs a geometric transformation which

can be used to shrink or zoom the size of an image.

 The shrink or Zoom, depends upon the vales of scaling factor.

 For values between 0 &1, the resultant Image is shrinked.

 For values greater than 1, the resultant Image is expanded by

that factor.
Scaling:
 The Scaling is easily accomplished by using the equations:

X2 = X1 * sx ===>(i)
Y2 = Y1 * sy ===>(ii)
Z2 = Z1 * sz ===>(iii)

 Where (X2, Y2, Z2) are new points and expressed in Matrix form as:

X2 sx 0 0 0 X1
V* = T
Y2 = 0 sy 0 0 Y1
V
Z2 0 0 sz 0 Z1
1 0 0 0 1 1

Where T represents the Scaling Matrix.

Scaling:
Original Image Scaled Image at 0.5*0.5
Scaling:
 The Initial Scale can be traced back by applying the Inverse
Scaling, that is:

V = T-1
V* X1 1/Sx 0 0 0 x2
y1 0 1/Sy 0 0 y2
z1 0 0 1/Sz 0
z2 1 0 0 0 1
1
1/Sx 0 0 0
0 1/Sy 0 0 Inverse Translation
T-1
0 0 1/Sz 0 Matrix
0 0 0 1

T . T-1 = I
Rotation :
 The transformation used for 3-D rotation is inherently more
complex than the Transformations discussed so far.

 The simplest from of these transformations is for Rotation of a

point about the coordinate axes.

 To rotate a point about another arbitrary Point in space, we

require three transformation:

 The first translate the arbitrary point to the Origin,

 The second performs the rotation, and
 The third translates the point back to its Original position.
Rotation :

 Rotation of a point about the Z coordinate axis by an angle  is

achieved by using the transformation.

 The rotation angle  is measured clockwise when looking at the

region from a Point on the +Z axis. This transformation affects
only the values of X and Y coordinates.

 The rotation operator performs a geometric transform which

maps the position ( x1, y1 ) of a picture element in an input
image onto a position ( x2, y2 ) in an output image by rotating
it through a user-specified angle ө about an origin (0,0).
 Rotation along Z-axis:
Y
P2(X2,Y2,Z2)
R is length of vector.
P1(X1,Y1,Z1)
φ is initial angle with X
axis.
R
Y2

Y1
R
θ
φ
0 X2 X

X1

Z
Rotation along Z-axis:
 Assume that the Vector (R) is making an angle of φ with X
axis.

 So initially, the components can be achieved as:

 X1=R cos φ
 Y1=R sin φ
 Z1=Z1

 If we rotate the vector R by θ around Z-axis in Counter

clock wise Direction, then now point formed will be given
by:

 X2=R cos (φ+θ)

 Y2=R sin (φ+θ )
 Z2=Z1 (Un changed axis)
Rotation along Z-axis:
 By expanding the terms of Sin and Cosine, we get:

X2 = R cosθ cosφ - R sinφ sinθ

Y2 = R sinθ cosφ + R sinφ cosθ
Z2= Z1

 But, we have assumed initially that:

R cosφ=x1 ===>(a)
R sinφ=y1 ===>(b)

 So our new point (x2,Y2,z2) becomes:

X2 = X1 cosθ - Y1 sinθ
Y2 = X1 sinθ + Y1 cosθ
z2= z1
Rotation along Z-axis:
 For matrix representation, these equations can be interpreted
as:

X2 = X1 cosθ - Y1 sinθ + 0
Y2 = X1 sinθ + Y1 cosθ + 0
z2 = 0 + 0 + z1

 In form of matrices, the rotation operation can be given as:

x2 cosθ -sinθ 0 x1
y2 sinθ cosθ 0 y1
z2 0 0 1 z1

V* = RθZ V
Inverse Rotation along Z-axis:

In form of matrices, the Inverse rotation operation can be

given as:

V = (RθZ)-1 V*
x1 cosθ sinθ 0 x2
y1 -sinθ cosθ 0 y2
z1 0 0 1 z2

(RθZ)-1 represents the Inverse Rotation Matrix.

(RθZ). (RθZ)-1 = I
Rotation along Y-Axis:
 Assume that the Vector (R) is making an angle of φ with Z-axis.

 So initially, the components can be achieved as:

 X1=R sin φ
 Y1=Y1
 Z1=R cos φ

 If we rotate the vector R by β around Y-axis in Counter clock

wise Direction, then now point formed will be given by:

 X2=R sin (φ+ β)

 Y2= Y1 (Un changed axis)
 Z2=R cos (φ+ β )
Rotation along Y-Axis:
 By expanding the terms of Sin and Cosine, we get:

X2 = R sinφ cos β + R cosφ sin β

Y2= Y1
Z2 = R cosφ cos β - R sinφ sin β

 But, we have assumed initially that:

R cosφ=x1 ===>(a)
R sinφ=z1 ===>(b)

 So our new point (x2,Y2,z2) becomes:

X2 = X1 cosβ + Z1 sinβ
Y2 = Y1
z2 = Z1 cosβ - X1 sinβ
Rotation along Y-Axis:
 For matrix representation, these equations can be interpreted
as:

X2 = X1 cosβ + 0 + Z1 sinβ
Y2 = 0 + Y1 + 0
z2 = -X1 sinβ + 0 + Z1 cosβ

 In form of matrices, the rotation operation can be given as:

x2 cosβ 0 +sinβ x1
=
y2 0 1 0 y1
z2 -sinβ 0 cosβ z1

V* = R β
Y
V
Inverse Rotation along Y-axis:

In form of matrices, the Inverse rotation operation can be

given as:
V = (R β Y)-1 V*

x1 cosβ 0 -sinβ x2
y1 0 1 0 y2
z1 sinβ 0 cosβ z2

(R β Y)-1 represents the Inverse Rotation Matrix.

(R β Y).(R β Y)-1 = I
Rotation along X-Axis:
 Assume that the Vector (R) is making an angle of φ with Y-axis.

 So initially, the components can be achieved as:

 X1=X1
 Y1=R cosφ
 Z1=R sinφ

 If we rotate the vector R by α around X-axis in Counter clock

wise Direction, then now point formed will be given by:

 X2=X1 (Un changed axis)

 Y2=R cos (φ+α )
 Z2=R sin (φ+α )
Rotation along X-Axis:
 By expanding the terms of Sin and Cosine, we get:

X2= X1
Y2 = R cosφ cos α - R sinφ sinα
Z2 = R sinφ cos α + R cosφ sin α

 But, we have assumed initially that:

R cosφ=y1 ===>(a)
R sinφ=z1 ===>(b)

 So our new point (x2,Y2,z2) becomes:

X2 = X1
Y2 = Y1 cos α - Z1 sinα
z2 = Z1 cos α + Y1 sin α
Rotation along X-Axis:
 For matrix representation, these equations can be interpreted
as:

X2 = X1 + 0 + 0
Y2 = 0 + Y1 cos α - Z1 sin α
Z2 = 0 + Y1 sin α + Z1 cos α

 In form of matrices, the rotation operation can be given as:

x2 1 0 0 x1
=
y2 0 cos α -sin α y1
z2 0 sin α cos α z1

V* = R α
X
V
Inverse Rotation along Y-axis:

In form of matrices, the Inverse rotation operation can be

given as:
V = (R α )
X -1
V*

x1 1 0 0 x2
y1 0 cosα sinα y2
z1 0 -sinα cosα z2

(R α )
X -1
represents the Inverse Rotation Matrix.

(R α
X
). (R α )
X -1
= I
Inverse Rotation:
The Inverse rotation of an angle, say (θo), around an axis is,
equivalent to the rotation by (–θo) around same axis.

For example, Inverse Rotation of θ around Z axis is equivalent to

the rotation of –θ around Z axis.

[RθZ]-1 = [R-θZ]
Cosθ sinθ 0 Cosθ sinθ 0
-sinθ cosθ 0 = -sinθ cosθ 0
0 0 1 0 0 1

Remember that Rotation Matrices are Ortho normal!

T 1 i  j
Ri  R j  
0 otherwise
Rotation:

Original Image
Rotation:

Rotated @ 180o Rotated @ 270o

Affine Transformation:

 The general affine transformation is commonly

written in homogeneous coordinates as shown
below:
Affine Transformation:
 Translation:
 By defining only the B matrix, this
transformation can carry out pure Translation:
Affine Transformation:
 Rotation:
 Pure Rotation uses the A matrix and is defined
as (for positive angles being clockwise
rotations):
Affine Transformation:
 Scaling
 Similarly, pure Scaling is: