Signals & Systems

MT - 252
Lecture # 14: Procedure for determining the coefficients of
Fourier Series Representation
Dr. Hafiz Zia Ur Rehman
hafizzia@mail.au.edu.pk

Department of Mechatronics Engineering

Air University, Islamabad
 Digital Image Processing, 3rd ed.
Gonzalez & Woods
www.ImageProcessingPlace.c
Procedure for determiningChapter
the coefficients
10
𝒂𝒏 om

❑ Assuming that a given Image

periodicSegmentation
signal can be represented with the series
of eq.
 2 
+ + jk  t
x (t ) =  k =
a e
k =−
jk0t
ae
k =−
k
 T0 
.

❑ For determining the coefficients 𝒂𝒌 , multiplying both sides of above eq.

by 𝒆−𝒋𝒏𝝎° 𝒕 ,
+
x ( t ) e − jn0t =  k e
a e
k =−
jk0t − jn0t

2
❑ Integrating both sides from 0 to 𝑻, T=
0
T T +

 ( ) dt = 
x t e − jn0t
 k e dt
a e jk0t − jn0t

0 0 k =−

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Gonzalez & Woods
www.ImageProcessingPlace.c
Procedure for determiningChapter
the coefficients
10
𝒂𝒏 om

❑ Interchanging the orderImage Segmentation

of integration and summation yield
T
 T + 
0 x ( t ) e dt = k (A)
j ( k − n )0t
− jn0t
ak   e dt .
=− 0 

❑ Rewriting the integral inside ‘[ ]’ using Euler's formula,

T T T
dt =  cos ( k − n ) 0t dt + j  sin ( k − n ) 0t dt.
j ( k − n )0t
e
0 0 0

❑ See that for 𝒌≠𝒏, both of the integrals on the right-hand side of eq. are
zero (i.e., area for 1 period is zero). For 𝒌 = 𝒏, the integrand on the left-
hand side of eq. equals 1, and thus, the integral equals T.
T
j ( k − n )0t T , k =n
e
0
dt = 
0, kn
,

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Gonzalez & Woods
T
j ( k − n )0t T , k =n
𝒂𝒏 0 e
www.ImageProcessingPlace.c
Procedure for determiningChapter
the coefficients om dt = 
10 0, kn
❑ Image
Consequently, the right-hand Segmentation
side of eq. reduces to 𝑻𝒂𝒏 . Therefore,
T
1
an =  x ( t ) e − jn0t dt
T 0
❑ This provides the equation for determining the coefficients.
❑ We were integrating over an interval of length T, which is an integral
number of periods of 𝒄𝒐𝒔(𝒌 − 𝒏)𝝎° 𝒕 and 𝒔𝒊𝒏(𝒌 − 𝒏)𝝎° 𝒕. Therefore,
we will obtain the same result if we integrate over any interval of length
T, i.e., if we denote integration over any interval of length T by ∫ , we
𝑇
have
T
j ( k − n )0t T , k =n
e dt = 
0, kn
,
0 The coefficient 𝑎𝑘 is the dc or constant
component of 𝒙(𝒕) and is given by eq:
1
an =  x ( t ) e − jn0t
dt a0 =
1
 x ( t ) dt
TT TT

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Gonzalez & Woods
www.ImageProcessingPlace.c
Procedure for determiningChapter
the coefficients
10
𝒂𝒏 om

❑ To summarize, if 𝒙(𝒕)Image
has a Fourier series representation [i.e., if it can
Segmentation
be expressed as a linear combination of harmonically related complex
exponentials], then the coefficients are given by eq.
T
1
an =  x ( t ) e − jn0t dt
T0
❑ This pair of equations (below) , then, defines the Fourier series of a
periodic continuous-time signal:
 2 
+ + jk  t
x (t ) =  k =
a e
k =−
jk0t
ae
k =−
k
 T0 
. Synthesis Equation

T T  2 
− jk  t
1 1
ak =  x ( t ) e − jk0t
dt =  x ( t ) e  T0 
dt Analysis equation
T0 T0
The set of coefficients 𝒂𝒌 are often called the Fourier series coefficients or
the spectral coefficients of 𝒙(𝒕).
© 1992–2008 R. C. Gonzalez & R. E. Woods
 Digital Image Processing, 3rd ed.
Gonzalez & Woods
www.ImageProcessingPlace.c
Example-04: Find the Fourier series coefficient
om 𝑎𝑘 for the following signal:
Chapter 10

x ( t ) = sin (0t )
Image Segmentation

Solution:- Expand 𝑥(𝑡) directly in terms of complex exponentials,

x ( t ) = sin (0t ) 
1 j0t − j0t
2j
e −e ( )
1 j (1)0t 1 j ( −1)0t
x (t ) = e − e
2j 2j Using
+

a1 =
1
−
j x (t ) = ae
k =−
k
jk0t
.
2j 2
1 j
a−1 = −  ak = 0, for k  1
2j 2

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Gonzalez & Woods
www.ImageProcessingPlace.c
Example-05: Find the Fourier series coefficient
om 𝑎𝑘 for the following signal:
Chapter 10
Image Segmentation  
x ( t ) = 1 + sin (0t ) + 2 cos (0t ) + cos  20t + 
 4
Solution:- Expand 𝑥(𝑡) directly in terms of complex exponentials,
  
j  20t +  − j  20t +  
 

x (t ) = 1 +
2j
e −e(
1 j0t − j0t
)1 j0t
2
(
+ 2  e + e − j0t +  e 
1
2  )4
+e  4



1  j ( −1) t  1 j  4   j ( 2) t  1 − j  4   − j ( 2) t
   
 1  j (1) t 
x ( t ) = 1 + 1 +  e + 1 −  e
0
+ e e
0
+ e e 0 0

 2j  2j    
2  2 
 
j 
    1 1
e 4
= cos   + j sin    +j
4 4 2 2
a0 = 1  
1 j  4  2
 1 
a1 = 1 +   1 −
j a2 = e  (1 + j ) ak = 0, k  0, 1, 2
2 4
 2j 2
 
 1  j
1 j  4 
a−2 = − e 
2
(1 − j ) or ak = 0, k  2
a−1 = 1 −   1 + 2 4
 2j 2

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Gonzalez & Woods
Magnitude and phase www.ImageProcessingPlace.c
of the Fourier coefficients of the signal
om
Chapter 10
a0 = 1
Image Segmentation
2
1 5
a1 = (1) +  
2

2 2 Even

1
a−2 =
2

0
a0 = tan −1    0
1
Odd
−1 
a1 = tan 
−1 2   −0.4636
 1 
 

 −1  
a−2 = tan −1    −
 1  4

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Gonzalez & Woods
www.ImageProcessingPlace.c
Example-06: Find Fourier series coefficients
om for the periodic square wave,
Chapter 10
sketched in Figure below and defined over one period as
Image Segmentation

1, t  T1

x (t ) =  T
0, T1 < t 
 2

Solution:- The given signal 𝒙(𝒕) is symmetry of about t = 0, it is convenient to

𝑻 𝑻
choose − , ≤ 𝒕 ≤ as the interval over which the integration is performed,
𝟐 𝟐
although any interval of length T is equally valid and thus will lead to the same
result.

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Gonzalez & Woods
For k = 0 www.ImageProcessingPlace.c
om
Chapter 10

1 Image Segmentation
x ( t ) dt
T1
a0 =
T  −T1

1 T1 2T1
a0 =  (1) dt 
T −T1 T
For general value of k
1 T1
ak =  x ( t ) e − jk0t dt
T −T1
jk0T1 − jk0T1
1 T1 −1 1 2  e − e 
ak =  (1) e − jk0t − jk0t T
dt 
1
e =  
T − T1 jk0 − T1 T k 0  2j 

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Gonzalez & Woods
www.ImageProcessingPlace.c
om
Chapter 10 jk0T1 − jk0T1
1 −1 1 2  e − e 
ak =  (1) e
T1 T
− jk0t − jk0t
dt 
Image Segmentation1
e =  
T − T1 jk0 − T1 T k 0  2j 

2sin ( k0T1 ) sin ( k0T1 )

ak =  , k  0.
k0T k
For T = 4T1
1
k = 1 → a1 = a−1 = ,

−1
k = 2 → a2 = a−2 = ,
3

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Gonzalez & Woods
www.ImageProcessingPlace.c
For T=4T1 om
Chapter 10
Image Segmentation

For T=6T1

For T=8T1

𝟐 𝐬𝐢𝐧 𝝎° 𝑻𝟏
The coefficients are regularly spaced samples of the envelope , where the
𝝎°
𝟐𝝅
spacing between samples, , decreases as T increases.
𝑻

© 1992–2008 R. C. Gonzalez & R. E. Woods

 References and Reading assignment
◼ The slides are prepared based on the following textbook:
◼ Chapter #3 (Art. 3.1-3.3), signals & systems by Alan V. Oppenheim, S. Hamid
Nawab, 2nd edition, Pearson education - prentice hall
◼ Reading Assignment : 3.1, 3.3, 3.3.2 with respective examples
◼ Special thanks to
◼ Prof. Dr. Muhammad Shehzad Hanif

13