INTRODUCTION

In 1820s an interesting geometry was discovered by Möbius in which transformations

of the plane map lines to lines, parallel lines to parallel lines, and preserve ratios of length
along lines and this geometry is called affine geometry,beacuse any two figures related by
such a transformation have a likeness or affinity to one another.
In Classical Greek geometry called Eucledian geometry after Euclid ,was regarded by
all in the early modern world as the true geometry of space .In Euclidean geometry, an
affine tranformation is a geometric transformation that preserves lines and parallelism
.More generally ,an affine transformation is an automorphism of an affine space ,that
is a fuction which maps an affine space onto itself while preserving both the dimension
of any affine subspaces (meaning that it sends points to points,lines to lines ,planes to
planes and so on) and the ratios of the lengths of parallel line segments. So the sets of
affine subspaces remain parallel after an affine transformation. An affine transformation
doesnot necessarily preserve angles between lines or distances between points, through it
does preserves ratios of distances between points lieing on straight line.Thus every linear
transformation is affine ,but not every affine transformation is linear transformation. Here
Geometrically ,affine transformations preserve collinearity,so they transform parallel line
into parallel lines and preserves ratios of diatances along parallel lines.
Thus in this section we goes through affine transformation and its fundamental proper-
ties.To put through it we met the idea that geometry can thought of in terms of a space
and a group acting on it is called Klenian view of geometry, After 19 th century theGerman
mathematician Felix Klein who proposed it first .Thus we examine Euclidean geometry
by Kleinian point of view.which gives the idea that Euclidean plane geometry uses the
space R2 ,and is concerned with those properties of figures that depend on the notion of
distance .The group associated with Euclidean geometry is the group of isometries of the
plane . The application behind affine transforamtion is image transformation

1
 PRELIMINARIES

Geometry and transformation

Definition 1. Geometry is the study of shape.The word comes from the
Greek word geo(Earth)and metria(measuring)
Definition 2. A Transformation in a geometry is a physical change to
location ,size or orientation of a given shape.The four types of transforma-
tions are rotation, translation, dilation or reflection.
Definition 3. Geometric transformation is any bijection of a set to
itself with some salient geometrical underpinning.
Definition 4. Rotation defined by turns a figure around a fixed point
,often the origin.and it doesn’t change the size and shape of the object.
Definition 5. Translation defined by sliding the figure a set distance
up,down,left or right and it doesn’t change the size and shape of the object.
Definition 6. Reflection defined by fliping the shape over a specified line
and size and shape are not changed.example x-axis ,y-axis and y=x.
Definition 7. An isometry of R2 is a function which maps R2 onto R2
and preserves distances.
Every isometry of R2 has one of the following forms:
ˆ a translation along a line in R2

ˆ a reflection in a line in R2

ˆ a rotation about a point in R2

ˆ a composite of translations,reflections and rotations in R2

The set of all isometries of R2 forms a group under composition of functions.

Remark 1. The composite of two isometries is an isometry
Example 1 (Every reflection is an isometry). Let 4ABC be a triangle in
which ∠ABC = ∠ACB.Prove that AB=BC.

2
Solution 1. First reflect the triangle in the perpendicular bisector of BC,
So that the points B and C changes places and the point A moves to some
0
point A , say. Since reflection preserves angle ,It follows that
0
∠A BC = ∠ACB.
0 0
So ∠A BC= ∠ABC.But this can happen only if A lies through A and B.
Similarly ,
0
∠A CB = ∠ABC = ∠ACB
0 0
.So A must also lie on the line through A and C.This means that A
and A must coincide.Hence the line segment AB reflects to the line seg-
ment AC,and vice versa .Since reflection preserves lengths, it follows that
AB=AC .
Problem 1. Let A and B be two points on a circle ,and let the tangents
to the circle A and B meet at P.Prove that AP=BP.
Solution 2. Consider a reflection in the line which passes through P and
the center of the circle.Under the reflection p remains fixed and the circle
0
maps onto itself.Then the point A maps to a point A on the circle, and
0
so the tangent PA maps onto the line PA . Now the tangent PA meets the
circle at a single point A,so the image of the tangent must meet the circle
0
at a single point.it happens only if A coincides with B. Hence the line
segment PA is reflected onto the line segment PB.Since reflection preserves
lengths,it follows that PA=PB.
Example 2 (A composition of translation, reflections, and rotation is an
isometry). Prove that if 4 ABC and 4DEF are two triangles such that
AB=DE,AC=DF and ∠ABC = ∠EDF,then BC=EF, ∠ABC= ∠DEF and
∠ACB= ∠DFE
Solution 3. To prove an isometry which maps 4ABC onto 4DEF.We
construct the isometry with the translation which maps A to D,shows the
0 0 0 0
translation maps 4ABC onto 4DB C where B and C are the images of
B and C under the translation .

graph
0
Since DF=AC gives the translation maps AC onto DC ,it follows that
0 0
DF=DC .We can therefore rotate the point C , about D unit it coincide

3
 0 0
with the point F.This rotation maps 4DB C onto 4DB ” F, where B ” is
0
the image of B under the rotation. shows that

∠F DE = ∠CAB(given)
0 0
= ∠C DB (translation)
= ∠F DB ” (rotation)

So either B ” lies on DE or the reflection of B ” in the line FD lies on DE.

Also

DB = AB(given)
= DB 0 (translation)
= DB ” (rotation)

It follows that either B ” coincides with E or the reflection of B ” in the line

FD coincides with E. So,composing translation ,rotation and reflection , we
obtaion the required isometry that maps 4 ABC and 4DEF .Since isom-
etry preserves length and angle , it follows that BC=EF, ∠ABC =∠DEF
and ∠ACB =∠DFE,
Definition 8. Euclidean geometry is the study of those properties of
figures that are unchanged by the group of isometries. These properties are
called Euclidean properties and include distance,angle collinearityof points
and concurrence of lines .
Definition 9. The kleinian view of geometry is the idea that geometry
can be thought of in terms of a group of transformations acting on a space
.
Definition 10. A Euclidean transformation of R2 is a function t : R2
→ R2 of the form
t(x) = Ux + a
where U is an orthogonal 2 × 2 matrix and a ∈ R2 .The set of all Euclidean
transformations of R2 is denoted by E(2).
Remark 2. A matrix U is orthogonal if U −1 = U T . That is , if U T U=I
This is equivalent to saying that the columns of U are orthonormal.
Theorem 1. Every isometry R2 is a Euclidean transformation of R2 and
viceversa

4
Proof. Consider the Euclidiean geometry has an algebraic representation
of R2 is given by
      
x cos θ − sin θ x e 2

t: 7→ + (x, y) ∈ R (1)
y sin θ cos θ y f
is an isometry . also it is the composite of an anticlockwise rotation through
an angle θ about the origin followed by a translation (e ,f) similarly the
function
      
x cos θ sin θ x e 2

t: 7→ + (x, y) ∈ R (2)
y sin θ − cos θ y f
is an isometry.Also it is the composite of a reflection in a line through the
origin that makes an angle θ/2 with the x -axis followed by a translation
through the vector (e,f). Now consider any isometry t can be written in
the form
t(x) = t0 (x) + (e, f ) x ∈ R2


(3)
where t0 is an isometry which fixes the origin. Now we let t0 be the trans-
formation defined by t0 (x)=t(x)-(e,f).This is an isometry because it is the
composite of the isometry t and translation through the vector -(e,f).It fixes
the origin t0 (0)=t(0)-(e,f)=0.Now the isometry that fixes the origin must
be either a rotation about the origin(1) or a reflection in a line through the
origin(2). we can represent any isometry by one or other forms given in
(1) and (2).So together (1) and (2) provide us an algebraic representation
of all possible isometries of the plane. Coversely,
If U is any orthogonal matrix , then its columns are orthonormal.In par-
ticular ,its first and second
 columns
 have unit
 length and can therefore
cos θ cos φ
be written in the form and respectively for some θ and
sin θ sin φ
φ.To prove orthonormal we must have cos θ · cos φ + sin θ · sin φ = 0,so that
tan θ · tan φ = −1 and Hence φ = θ ± π2 .It follows that second column must
be
cos(θ + π2 )
   
− sin θ
= or
sin(θ + π2 ) cos θ
cos(θ − π2 )
   
sin θ
=
sin(θ − π2 ) − cos θ
So    
cos θ − sin θ cos θ sin θ
U= or U=
sin θ cos θ sin θ − cos θ

5
It follows that every Euclidean transformation t(x)=Ux+a of R2 has one
of the form given in 91) and(2).

Theorem 2. The set of all Euclidean transformations of R2 forms a group

under the operation of composition of functions.
Proof. let we first consider closure. Suppose t1 and t2 are to Euclidean
transformations given by

t1 (x) = U1 x + a1 and t2 (x) = U2 x + a2

where U1 and U2 are orthogonal 2 × 2 matrices. Then the composite t1 ◦t2

is given by

t1 ◦ t2 (x) = t1 (U2 x + a2 )
= U1 (U2 x + a2 ) + a1
= U1 U2 x + (U1 a2 + a1 )

This is a Euclidean transformation since U1 U2 is orthogonal.

(U1 U2 )T = U2T U1T = U2−1 U1−1 = (U1 U2 )−1

(since(AB)T = B T AT )
so the set of Euclidean transformation is closed under composition of func-
tions. Now consider That under composition of functions the identity is
the transformation given by
i(x) = x
. This is a Euclidean transformation since it can be written in the form

i(x) = Ix + 0
where I is the 2 × 2 identity matrix is orthogonal
Now consider the inverse of Euclidean transformation

t1 (x) = U1 x + a1 is given by t2 (x) = U2 x + a2

and also composition of function is always associative .so all four group
properties hold.

6
Definition 11. The inverse of the Euclidean transformation

t(x) = U x + a

is given by
t−1 (x) = U −1 x − U −1 a
Example 3. Consider two triangle 4ABC and 4 DEF such that AB=DE,AC=DF
and ∠BAC =∠EDF, then there is a Euclidean transformation which maps
4ABC and 4 DEF
Solution 4. The existence of transformation enable us to deduce that both
triangles have same Euclidean properties.Also able to deduce that BC=EF,∠ABC=∠DE
and ∠ACB=∠DEF.

graph

Definition 12. A figure F1 is Euclidean congruent to a figure F2 if

there is a Euclidean transformation which maps F1 onto F2.In general two
figures are congruent if they have same size and shape .
Example 4. Any two circles of unit radius are Euclidean -congruent to a
each other.which proves that translation makes their center coincides.Similarly
The set of all line segments of length 1 and The set of all squares that have
sides of length 2.
Theorem 3. Euclidean congruence is an equivallence relation .
Proof. To prove equivallence relation,
we have to prove reflexive,symmetric and transitive axioms. First to show
reflective ,by using the existence of an identity transformation.For all fig-
ures F in R2 , the identity maps F onto itself, so Euclidean -congruence is
reflexive.
Now to show symmetric,by using the existence of inverse transformation.Let
a figure F1 in R2 be congruent to a figure F2 ,and let t be a Eucldean trans-
formation which maps F1 to F2 .Then the inverse Euclidean transformation
t−1 maps F2 onto F1 .Thus Euclidean -congruence is symmetric and To
prove the transitive,by using the closure axiom for the group of Euclidean
transformation.Let a figure F1 in R2 be congruent to a figure F2 ,and let
F2 congruent to a figure F3 .Then there exist a euclidean transformation
t1 maps F1 onto F2 and t2 maps F2 onto F3 , so that F1 is congruent to
F3 .Hence Euclidean congruence is translative.

7
Definition 13. A figure F1 is G congruent to a figure F2 in some ge-
ometry defined by a group G of transformation in G which maps F1 onto
F2 .
Remark 3. G congruence is an equivallence relation.

8
 Affine transformations and parallel projections
Chapter 1

1 Affine transformations
Definition 14. An Affine transformation of R2 is a function t : R2
→ R2 of the form

t(x) = Ax + b,
where A is an invertible 2 × 2 marix and b ∈ R2 is denoted by A(2)
Remark 4. Every Eucliedean transformation of R2 is an Affine trans-
formation of R2 . Since every orthogonal matrix is invertible .(In terms of
groups, the group of Euclidean transformations of R2 is a proper subgroup
of the group of affine transformation of R2 .This means that all properties
of figures that are preserved by affine transformations must be preserved
also by Euclidean transformations).
Problem 2. Determine whether the following transformations of R2 is an
affine transformation.
       
1 3 4 −6 5 2
(a)t1 (x) = x+ (b)t2 (x) = x+
1 2 −2 3 2 1
       
−2 −1 1 5 −3 0
(c)t3 (x) = x+ (d)t4 (x) = x+
8 4 3 −2 2 0
Solution 5. (a) Compare the given transformation with the general form
We have  
1 3
t(x) = Ax + b, where A =
1 2
We know that it is a affine transformation ,if A is invertible. To check A
is invertible
 ,find
 the determinant of A
1 3 
That is  =-1 6= 0.Thus A is invertible and form t1 is an affine trans-
1 2
formation .
(b)Similarly to find t2 is an affine transformation,find the determinant
of A  
−6 5
That is  = -276= 0.Thus A is invertible and form t2 is an affine trans-
3 2
formation.

9
 (c)Tofind t3 is
 an affine transformation,find the determinant of A
−2 −1
That is  =0.Thus A is not invertible and do not form an affine
8 4
transformation and (d)to find t4 is an affine transformation,find the deter-
minant of A 
 5 −3
That is  =4 6= 0.Thus A is invertible and form t4 is an affine trans-
−2 2 
formation.
Problem 3. State whether or not composite is an affine transformation.
(a)t1 ◦ t2 (b)t2 ◦ t4
Solution 6. (a) To find t1 ◦ t2 Consider
      
1 3 −6 5 2 4
t1 (t2 (x)) = x+ +
1 2 3 2 1 −2
       
1 3 −6 5 1 3 2 4
= x+ +
1 2 3 2 1 2 1 −2
   
3 11 9
= x+
0 9 2
Which is general form of affine transformstion
  by comparing. Thus to
3 11
check take the determinant of A,Since   =27 6= 0.
0 9
Similarly (b)To find t2 ◦ t4 ,consider
     
−6 5 5 −3 2
t2 (t4 (x)) = x +
3 2 −2 2 1
   
−40 28 2
= (x) +
15 −5 1
 
−40 28 
.Since   =-108 6= 0 .Therefore t2 ◦ t4 form an affine transforma-
15 −5
tion.
Theorem 4. The set of affine transformation A(2) forms a group under
the operation of compositon of functions.
Proof. Let t1 and t2 be affine transformations given by

t1 (x) = A1 x + b1 and t2 (x) = A2 x + b

10
where A1 and A2 are invertible 2 × 2 matrices. Then, for each x ∈ to R2

(t1 ◦ t2 )(x) = t1 (A2 x + b2 )

= A1 (A2 x + b2 ) + b1
= A1 A2 x + (A1 b2 ) + b1 )
since A1 and A2 are invertible , it follows that A1 A2
is also invertible. so by definition t1 ◦ t2 is an affine transformation.Thus
we get affine transformation is closed under composition of functions.
Now let i be the affine transformation given by
i(x) = Ix + 0 (x ∈ R2 ),
where I is the identity matrix .If t is an affine tansformation given by
t(x) = Ax + b (x ∈ R2 ),
then ,for each x ∈ R2

(t ◦ i)(x) = A(Ix + 0) + b = Ax + b = t(x).

and
(i ◦ t)(x) = I(Ax + 0) + b = Ax + b = t(x).
Thus t ◦ i and i ◦ t =t . Hence i is the identity transformation. Now , If t
is an arbitrary affine transformation given by t(x)=Ax+b (x belong to R2
0
) Then we can define another affine transformation t by
0
t (x) = A−1 x − A−1 b.
Now for each x belong to R2 , we have
0
(t ◦ t ) = t(A−1 x − A−1 b)
= A(A−1 x − A−1 b) + b
= (AA−1 x − AA−1 b) + b
= (x − b) + b
=x
Also
0 0
(t ◦ t) = t (Ax + b)
= A−1 (Ax + b) − A−1 b
= (A−1 Ax + A−1 b) − A−1 b
= (x + A−1 b) − A−1 b
=x

11
 0 0 0
Thus t ◦ t = t ◦ t = i. Hence t is an inverse for t.
and composition of functions is always associative . thus the four group
axioms satisfies that the set of affine transformation A(2) forms a group
under composition of functions.

Definition 15. The inverse of the affine transformation

t(x) = Ax + b
is given by
t−1 (x) = A−1 x − A−1 b
Definition 16. Affine geometry is the study of those properties (Affine
properties ) of figures in the plane R2 that are preserved by affine trans-
formations.
Problem 4. Find the inverse of the affine transformation
   
1 3 4
t(x) = x+
1 2 −2
Solution 7. First find
 
1 3
A−1 =
1 2
 
1 d −b
=
ad − bc −c a
 
1 2 −3
=
2 − 3 −1 1
 
2 −3
= (−1)
−1 1
 
−2 3
=
1 −1

Therefore
     
−2 3 x1 −2 3 4
t−1 (x) = −
1 −1 x2 1 −1 −2
   
−2x1 + 3x2 −14
= −
x1 − x2 6

12
It follows that

x = −2x1 + 3x2 + 14 and y = x1 − x2 − 6

Basic properties of Affine transformation

ˆ maps straight lines to straight line

ˆ map parallel straight lines to parallel straight lines

ˆ preserves ratios of lengths along a given straight line

13
 chapter 2

2 Parallel projection
Definition 17. A Parallel projecttion is a one -one mapping of R2 is de-
fined in the following way.Let π1 and π2 be planes in R3 , with parallel rays
of light shining through them,Then the function p which maps each points
0
p in π1 to the corresponding point p in π2 is a parallel projection from π1
onto π2 .

graph

Remark 5. If π1 and π2 are parallel , then the parallel projection from π1

onto π2 is an isometry.

Basic properties of Parallel projection

ˆ maps straight lines to straight line

ˆ map parallel straight lines to parallel straight lines

ˆ preserves ratios of lengths along a given straight line

Property 1. A parallel projection maps straight lines to straight lines.

Proof Let l be a line in the plane π1 ,and let p be a parallel mapping
π1 onto the plane π2 .Now consider all the rays associated with p that pass
through l,Since these rays are parallel , they must fill a plane .Call this
plane π

graph

The image of l under p consists of those points where the rays that pass
through l meet π2 .But these points are simply the points of intersection of
π with π2 .Since any two intersecting planes in R3 meet in a line , it follows
that the image of l under p is a straight line . proof
Property 2. A parallel projection maps parallel straight lines to parallel
straight lines ProofLet l1 and m1 be parallel lines in the plane π1 , and let p
be a parallel projection mapping π1 onto the plane π2 .Let l2 and m2 be the

14
lines in π2 that are images under p of l1 and m1 .

graph

If l2 and m2 are not parallel ,they meet at some point, P2 say.Let P1 be

the points of π1 which maps to π2 .Then P1 must lie on bothl1 and m1 .Since
l1 and m1 are parallel,no such point of intersection can exist ,which is a
contradiction .It follows that l2 and m2 must indeed be parallel. proof
Property 3. A parallel projection preserves ratios of lengths along a given
straight line
ProofLet A,B,C be three points on a line in the plane π1 ,and let p be a
parallel projection mapping π1 onto the plane π2 .Let P,Q,R be the points
in π2 that are the images under p of A,B,C .

graph

We know from Property 1 that P, Q ,R lie on the line ;we have to show
that the ratio AB : AC is equal to the ratio PQ:PR.
If the planes π1 and π2 are parallel ,then the parallel projection p is an
isometry, and so the ratios AB: AC and PQ:PR are equal ,as required .On
the other hand , if π1 and π2 are not parallel , then we can construct a
plane π through the point P which is parallel to π1 ,as shown in the figure

graph
0
This plane intersects the ray through B and Q at some points B , and
0
the ray through C and R at some points C . So in this case the ratios
0
AB:AC and PB :PC 0 are equal.
0 0 0
Now consider 4PC R. The lines B Q and C R are parallel ,since they are
0 0
rays from parallel projection. Hence B Q meets the sides PR and PC in
0 0
equal ratios. Thus PQ : PR = PB : P . It follows that PQ:PR=AB:AC,are
required. proof

15
3 Affine transformations and parallel projections
3.1 Relationship between affine transformation and parallel pro-
jection
Theorem 5. Each Parallel projection is an Affine transformation
Proof. Case:(1)First we consider a parallel projection p of plane π1 onto a
plane π2 .

graph

For the moment suppose that the planes are aligned so that the origin in
π1 is mapped to the origin in π2 since ratios of length are preserved along
a straight line , we must have for any vector v ∈ R2 and any λ ∈ R2 ,

p(λv) = λp(v) (4)

Now let v and w be two position vectors in π1 . Their sum , v+w , is
found from the parallelogram law for addition of vectors , as shown in the
diagram below. The image under p in π2 are p(v) and p(w) and the sum
of these two vectors is p(v)+p(w), But a parallel projection maps parallel
lines onto parallel lines .so it must map parallelograms onto parallelograms.
Hence it must map the parallelogram in pi 1 onto the parallelogram in π2 .
Therefore the map v+w to p(v)+p(w) is defined as

p(v + w) = p(v) + p(w) (5)

it follows from 1 and 2 that p must be linear transformation of R2 onto

itself. Hence there exist some matrix A such that v ∈ R2

p(v) = Av

since the linear transformation p is invertible , it follows that a is in-

vertible . Now suppose that the parallel projection maps the origin in π1
to some point in B with position vector b in π2 .if we temperorily construct
a new set axes in π2 that are parallel to the original axes and is interscet at
the point B .Then for some invertible 2 × 2 matrix A ,we obtain

p(v) = Av + b

which is of the form Affine transformation Therefore p must be an Affine

transformation.

16
4 Remark
The coverse of the theorem is not true that every affine transformation can
be represented as a parallel projection(that is An affine transformation is
not necessarily a parallel projection)
Theorem 6. An Affine transformation can be expressed as the composite
of two parallel projections.
Proof. Consider an affine transformation t: R2 → R2 has the form

t(x) = Ax + b (x ∈ R2 ) (6)

; where A is an invertible 2×2 matrix . Now t is not a linear transformation

unless b=0,but we can use similar method to find linear transformations
to determine A and b. from equation t(0)=b,so b is the image of the origin
under t.let e and f are coordinates of t(0) that is,
   
a b e
A= and b =
c d f
where a,b,c,d are real numbers that have yet to be found.It follows that
from equation (5) that the images under t of the points (1,0) and (0,1) are
given by
        
a b 1 e a e
+ = +
c d 0 f c f
and
        
a b 0 e b e
+ = +
c d 1 f d f

so t(0)=(e,f),we know the points onto which (1,0)and(0,1) are mapped by

t,then we can determine the values of a,b,c,and d . then

(a, c) = t(1, 0) − (e, f ) and (b, d) = t(0, 1) − (e, f )

that is an affine transformation t, the images t(0,1),t(1,0)and t(0,0) = (e,f)

cannot be collinear ,for if they were , then (a,c) and (b,d) would be linearly
depended and A is noninvertible. It follows that an affine transformation
is uniquely determined by the non collinear points (0,0)(1,0) and (0,1).

17
 chapter 3

5 Properties of Affine Transformation

5.1 image of sets under affine transformation
Theorem 7. Determine the image of a line or conic in R2 under an affine
transformation .
the image of a line or conic in R2 under an Affine transformation of R2
is a function t : R2 → R2 is given by

t(x) = Ax + b (7)

,where a is an invertible 2 × 2 matrix and b ∈ R2 .

The set of such transformations forms a group, in which the transfor-
mation inverse to t is given by

t−1 (x) = A−1 x–A−1 b (8)

Let x and coordinates (x,y) denote the points in the domin of R2 , and x’
and coordinates (x’,y’) denote points in the codomain . from equation 1
and 2 we get the equation of the form

0
x = Ax + b (9)
0
x = A−1 x − A−1 b (10)

Example 5. Determine the image of the line y=2x under the affine trans-
formation    
4 1 2
t(x) = x+ (x ∈ R2 ) (11)
2 1 −1
0 0
Solution 8. Let (x,y) be an arbitary point on the line y =2x , and let (x ,y )
be the image under t,Then
 0     
x 4 1 x 2
0 = +
y 2 1 y −1
Next we use the equation
0
x = A−1 x − A−1 b (12)

18
 0 0
to express (x, y) in terms of (x , y ). We have
 −1  1 1
  1 1
   3 
4 1 − − 2
A−1 = = 2 2 and A−1 b = 2 2 = 2
2 1 −1 2 −1 2 −1 −4
so    1   0  3
1
x − x −2
= 2 2 0 +
y −1 2 y 4
It follows that under the inverse mapping t−1 we have
1 0 1 0 3 0 0
x= x − y − and y = −x + 2y + 4
2 2 2
0 0
Since x and y are related by equation y=2x, it follows tht x and y are
related by the equation
0 0 0 0
−x + 2y + 4 = 2 21 x − 21 y − 32

, Which simplifies to
2x0 − 3y 0 = 7
Dropping the dashes , we see that the image of the line y=2x under t is the
line
2x − 3y = 7
.
Example 6. Determine the image of the circle x2 + y 2 = 1 under affine
transformation.
 1 1
  3  
− − 2
t(x) = 2 2 x+ 2 x∈R (13)
−1 2 4
Solution 9. Let (x,y) be an arbitary point on the circle x2 + y 2 =1 , and
0 0
let (x ,y ) be the image under t,Then
 0  1 1
    3
x − x −2
0 = 2 2 +
y −1 2 y 4
Next we use the equation
0
x = A−1 x − A−1 b (14)
0 0
to express (x, y) in terms of (x , y ). We have
1 −1
 1       3  
−1 − 4 1 −1 4 1 −2 −2
A = 2 2 = and A b = =
−1 2 2 1 2 1 4 1

19
so      0  
x 4 1 x 2
= 0 +
y 2 1 y −1
It follows that under the inverse mapping t−1 we have
0 0 0 0
x = 4x + y + 2 and y = 2x + y − 1
0 0
Since x and y are related by equation y=2x, it follows tht x and y are
related by the equation
0 0
2 0 0
2
4x + y + 2 + 2x + y − 1 = 1
0 0
20x + 2y + 2 = 0
0 0
x +y =1

Dropping the dashes ,we see that the image of the circle x2 + y 2 = 1
under t is the circle x2 + y 2 = 1.

20
6 The Fundamental Theorem of Affine Geometry
0 0 0
Let p ,q,r and p , q , r be two sets of three non-collinear points in R2 .Then
0 0 0
there is a affine transformation t which maps p,q and r to p , q , r respec-
tively. and the affine transformation is unique.
Property 4. Determine the unique affine transformation t(x)=Ax+b which
maps (0, 0), (1, 0) and (0, 1) to the three non collinear points p,q and re-
spectively
1. take b=p;
2. take A to be the matrix with columns given by q-p and r-p.
Example 7. Determine the affine transformation t which maps the points
(0, 0), (1, 0) and (0, 1) to the points (3, 2), (5, 8) and (7, 3) respectively
Solution 10. let t be the affine transformation given by
      
x a b x e
t: 7→ + (15)
y c d y f
which maps t(0, 0) to (3, 2), t(1, 0) to (5, 8), t(0, 1) to (7, 3). Since t(0, 0) =
(3, 2) follows (6) that e = 3 and f = 2. t(1, 0) = (5, 8)so that

          
5 a b 1 3 a 3
= + = +
8 c d 0 2 c 2
The first column of the matrix for t is therefore
       
a 5 3 2
= − =
c 8 2 6
Finally t(0,1)=(7,3) ,so that
          
7 a b 0 3 b 3
= + = +
3 c d 1 2 d 2
The second column of the matrix for t is therefore
       
b 7 3 4
= − =
d 3 2 1
Hence the desired affine transformation is given by
      
x 2 4 x 3
t: 7→ +
y 6 1 y 2

21
Theorem 8. Determine the affine transformation t which maps three non
0 0 0
collinear points p,q ,r to another three non-collinear points p , q , r respectively.
Proof.
Example 8. Determine the affine transformation which maps the points
(2, 3), (1, 6) and (3, −1)to the points (1, −2), (2, 1) and (−3, 5) respectively.
Solution 11. Let the affine transformation t1 which maps the points (0, 0), (0, 1)
and (0, 1) to the points (2, 3),(1, 6) and (3, −1) respectively, is gived by
   
−1 1 2
t1 = x+
3 −4 3

and also let the affine transformation t1 which maps the points (0, 0), (0, 1)
and (0, 1) to the points (1, −2), (2, 1) and (−3, 5) respectively, is gived by
   
1 4 2
t2 = x+
3 1 −2
To find inverse of the affine transformation t1 which maps the points
(0,0),(0,1) and (0,1) to the points (2,3) ,(1,6) and(3,-1) respectively, is
gived by    
−1 1 2
t1 = x+
3 −4 3
To find inverse of t1 is

t−1 −1 −1
1 (x) = A x − A b

fisrt compute
 −1  
−1 −1 1 −4 −1
A = =
3 −4 −3 −1
and     
−4 −1 2 −11
=
−3 −1 3 −9
so the inverse of t1 is given by
   
−1 −4 −1 −11
t1 (x) = x−
−3 −1 −9

22
Thus the affine transformation which maps the points (2,3),(1,6) and (3,-1)
to the points (1,-2),(2,1) ans (-3,5) respectively ,is given by

t(x) = t2 ◦ t−1 (x)

 1   
−4 −1 11
= t2 x+
−3 −1 9
      
1 −4 −4 −1 11 1
= x+ +
3 7 −3 −1 9 −2
     
8 3 −25 1
= x+ +
−33 −10 96 −2
   
8 3 −24
= x+
−33 −10 94
Definition 18. Two figures are affine -congruent if there is an affine trans-
formation which maps one onto the other.
Remark 6. All triangles are affine- congruent.
Theorem 9. An affine transformation maps straight lines to straight lines.
Proof. Let l be a line through a point with position vector p , and let the
direction of l be that of some vector a . Then

l = {p + λa : λ ∈ R}

Now let t : R2 → R2 be an affine transformation given by

t(x) = Ax + b
To find the image t of an arbitrary point p +λa on l as follows:

t(p + λa) = A(p + λa) + b

= (Ap + b) + λAa
= t(p + λAa

so the image of l is the set t

t(p) = {t(p) + λAa : λ ∈ R}

which is the line through t(p) in the direction of the vectorAa.

Theorem 10. An affine transformation maps parallel straight lines to par-
allel straight lines.

23
Proof. Let l1 and l2 be parallel lines through the points with position vec-
tors p and q respectively ,and let the direction of the lines be thst of the
vcetor a. Then
l1 = {p + λa : λ ∈ R} and l2 = {q + λa : λ ∈ R}
Let t:R2 → R2 be an affine transformation given by

t(x) = Ax + b
Then the images of l1 and l2 under the affine transformation are the sets
t(l1 ) = {t(p) + λAa : λ ∈ R} and t(l2 ) = {t(q) + λAa : λ ∈ R}
. These sets are straight lines which pass through the image point t(p)
and t(q), both in same direction as that of the vector Aa. Hence the two
image lines under t are parallel .
Theorem 11. An affine transformation preserves ratio of lengths along
parallel straight lines.
Proof. Consider the line segment under the affine transformation ,

graph

Let l be a line through a point with position vector p,and let the direc-
tion of l be that of some unit vector a.Then

t(l) = {p + λa : λ ∈ R}
Then the image of l under the affine transformation
t(x) = Ax + b
is the line
t(l) = {t(p) + λAa : λ ∈ R}
Now consider a segment l with endpoints p+λ1 a and p+λ2 a. Since a is a
unit vector , the length of the segment is
k(p + λ2 a) − (p + λ1 a)k = |λ2 − λ1 | · kak = |λ2 − λ1 |.
The image of the segment has endpoints t(p)+λ1 Aa and t(p)+λ2 Aa,so
the image of the segment has the length
k(t(p) + λ2 a) − (t(p) + λ1 a)k = |λ2 − λ1 | · kAak.

24
So that maps segments along l to segment along t(l),lengths are stretched
by the factor kAak.Since this factor is the same for all segments which lie
along lines parallel to a,it follows that the ratios of lengths along parallel
lines are unchanged by t.

25
7 Applications
Definition 19. Let 4ABC in the plane and join the midpoint of each side
of the triangle to the opposite vertex theselines are called the medians.
Theorem 12 (The Median Theorem). The medians of any triangle are
concurrent.
Proof. Consider a equilateral 4ABC in the plane,with AP,BQ and CR.Since
4ABC has sides of equal length,it must be symmetric about the line
AP.Thus the point at which BQmeets CR must be symmetrically placed
with respect to the line .Thus the lines AP,BQ and CR are concurrent if
the triangle is equilateral. .

26