You are on page 1of 29

Lecture 01

om
.c
ng
co
Digital Imaging Fundamentals

an
th
o ng
du
u
cu

CuuDuongThanCong.com https://fb.com/tailieudientucntt
Contents

This lecture will cover:

om
 Light and the electromagnetic spectrum

.c
 Image representation

ng
 Image sensing and acquisition

co
 Sampling, quantisation and resolution

an
th
o ng
du
u
cu

Digital Image Fundamentals 2


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Light And The Electromagnetic Spectrum

 Light is just a particular part of the electromagnetic

om
spectrum that can be sensed by the human eye.

.c
 The electromagnetic spectrum is split up according to

ng
the wavelengths of different forms of energy

co
an
th
o ng
du
u
cu

Digital Image Fundamentals 3


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Reflected Light

 The colours that we perceive are determined by the

om
nature of the light reflected from an object

.c
 For example, if white

ng
light is shone onto a

co
green object most

an
wavelengths are

th
absorbed, while green ng Colours
light is reflected from Absorbed
o

the object.
du
u
cu

Digital Image Fundamentals 4


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Sampling, Quantisation And Resolution

In the following slides we will consider what is involved in

om
capturing a digital image of a real-world scene

.c
 Image sensing and representation

ng
 Sampling and quantisation

co
 Resolution

an
th
o ng
du
u
cu

Digital Image Fundamentals 5


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Image Acquisition

 Images are typically generated by illuminating a scene

om
and absorbing the energy reflected by the objects in

.c
that scene.
 Typical notions of

ng
co
illumination and

an
scene can be way off:
th
ng
• X-rays of a skeleton
o
• Ultrasound of an
du

unborn baby
u
cu

• Electro-microscopic
images of molecules

Digital Image Fundamentals 6


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Image Sensing

 Incoming energy lands on a sensor material

om
responsive to that type of energy and this generates a

.c
voltage
 Collections of sensors are arranged to capture images

ng
co
an
th
o ng
du
u
cu

Imaging Sensor

Line of Image Sensors


Array of Image Sensors

Digital Image Fundamentals 7


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Image Sampling And Quantisation

 A digital sensor can only measure a limited number of

om
samples at a discrete set of energy levels

.c
 Quantisation is the process of converting a continuous

ng
analogue signal into a digital representation of this

co
signal

an
th
o ng
du
u
cu

Digital Image Fundamentals 8


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Image Sampling And Quantisation (cont…)

 Remember that a digital image is always only an

om
approximation of a real world scene

.c
ng
co
an
th
o ng
du
u
cu

Digital Image Fundamentals 9


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Image Representation

 Before we discuss image acquisition recall that a


digital image is composed of M rows and N columns of

om
pixels each storing a value.

.c
 Pixel values are most col

ng
often grey levels in the

co
range 0-255(black-white).

an
 We will see later on
that images can easily
th
ng
be represented as
o
du

matrices.
u
cu

f (row, col)
row

Digital Image Fundamentals 10


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Image Representation

om
.c
ng
co
an
th
o ng
du
u
cu

Digital Image Fundamentals 11


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Spatial Resolution

 The spatial resolution of an image is determined by

om
how sampling was carried out

.c
 Spatial resolution simply refers to the smallest

ng
discernable detail in an image

co
• Vision specialists will

an
often talk about pixel

th
size ng
• Graphic designers will
o
du

talk about dots per


inch (DPI)
u
cu

Digital Image Fundamentals 12


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Spatial Resolution (cont…)

om
.c
ng
co
an
th
o ng
du
u
cu

Digital Image Fundamentals 13


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Spatial Resolution (cont…)

om
.c
ng
co
an
th
o ng
du
u
cu

Digital Image Fundamentals 14


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Spatial Resolution (cont…)

om
.c
ng
co
an
th
o ng
du
u
cu

Digital Image Fundamentals 15


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Spatial Resolution (cont…)

om
.c
ng
co
an
th
o ng
du
u
cu

Digital Image Fundamentals 16


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Spatial Resolution (cont…)

om
.c
ng
co
an
th
o ng
du
u
cu

Digital Image Fundamentals 17


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Spatial Resolution (cont…)

om
.c
ng
co
an
th
o ng
du
u
cu

Digital Image Fundamentals 18


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Intensity Level Resolution

 Intensity level resolution refers to the number of

om
intensity levels used to represent the image

.c
• The more intensity levels used, the finer the level of

ng
detail discernable in an image

co
• Intensity level resolution is usually given in terms of the

an
number of bits used to store each intensity level

th
Number of Intensity
ng
Number of Bits Examples
Levels
o
du

1 2 0, 1
2 4 00, 01, 10, 11
u
cu

4 16 0000, 0101, 1111


8 256 00110011, 01010101
16 65,536 1010101010101010

Digital Image Fundamentals 19


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Intensity Level Resolution (cont…)
256 grey levels (8 bits per pixel) 128 grey levels (7 bpp) 64 grey levels (6 bpp) 32 grey levels (5 bpp)

om
.c
ng
co
an
th
o ng
du
u
cu

16 grey levels (4 bpp) 8 grey levels (3 bpp) 4 grey levels (2 bpp) 2 grey levels (1 bpp)

Digital Image Fundamentals 20


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Saturation & Noise

om
.c
ng
co
an
th
o ng
du
u
cu

Digital Image Fundamentals 21


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Resolution: How Much Is Enough?

The big question with resolution is always how much is


enough?

om
 This all depends on what is in the image and what you

.c
would like to do with it

ng
 Key questions include

co
an
• Does the image look aesthetically pleasing?

th
• Can you see what you need to see within the image?
ng
o
du
u
cu

Digital Image Fundamentals 22


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Resolution: How Much Is Enough? (cont…)

om
.c
ng
co
an
th
o ng
du

 The picture on the right is fine for counting the number


u
cu

of cars, but not for reading the number plate

Digital Image Fundamentals 23


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Relationships Between Pixels

 Neighbors of a pixel: consider a pixel p at (x,y)

om
.c
p

ng
co
an
 4-neighbors of p, N4(p): neighbors at (x+1,y), (x-
th
ng
1,y), (x,y+1), (x,y-1).
o

 4-diagonal neighbors of p, ND(p): (x+1,y+1), (x+1,y-


du

1), (x-1,y+1), (x-1,y-1).


u
cu

 8-neighbors of p, N8(p): N4(p) and ND(p).

Digital Image Fundamentals 24


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Relationships Between Pixels

 Connectivity: Two pixels are connected if they are


adjacent (e.g. 4-neighbors) and their gray levels

om
satisfy specified criterion of similarity (e.g. they are

.c
equal). Let V set of gray levels used to define

ng
co
connectivity (e.g. V={1}: binary image;
V={32,33,…,63,64}: gray-scale image).

an
th
• 4-connectivity (4-adjacent): p and q are 4-connected if
ng
their values from V and q is in N4(p).
o
du

• 8-connectivity (8-adjacent): p and q are 8-connected if


their values from V and q is in N8(p).
u
cu

Digital Image Fundamentals 25


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Relationships Between Pixels

 m-connectivity (m-adjacent, mixed connectivity): p


and q are m-connected if their values from V and

om
 q is in N4(p), or

.c
 q is in ND(p) and set S=N4(p)  N4(q) is empty (pixels  S are

ng
4-neighbors of both p and q, and whose values are from V).

co
 m-connectivity introduced to eliminate ambiguity in

an
path connections (resulted from allowing 8-
connectivity)
th
o ng
du
u
cu

Digital Image Fundamentals 26


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Relationships Between Pixels
 Path
• A (digital) path (or curve) from pixel p with coordinates

om
(x0, y0) to pixel q with coordinates (xn, yn) is a sequence

.c
of distinct pixels with coordinates:

ng
(x0, y0), (x1, y1), …, (xn, yn)

co
where (xi, yi) and (xi-1, yi-1) are adjacent for 1 ≤ i ≤ n.

an
th
Here n is the length of the path. ng
• If (x0, y0) = (xn, yn), the path is closed path.
o
du

• We can define 4-, 8-, and m-paths based on the type of


u
cu

connectivity used.

Digital Image Fundamentals 27


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Relationships Between Pixels

Examples: Connectivity and Path, V = {1, 2}

om
.c
ng
co
an
th
o ng
du
u
cu

Digital Image Fundamentals 28


CuuDuongThanCong.com https://fb.com/tailieudientucntt
Summary

We have looked at:

om

.c
Light and the electromagnetic spectrum

ng
 Image representation

co
 Image sensing and acquisition

an
 Sampling, quantisation and resolution
th
o ng
Next time we start to look at techniques for image
du

enhancement
u
cu

Digital Image Fundamentals 29


CuuDuongThanCong.com https://fb.com/tailieudientucntt

You might also like