Image processing

23/4/12 Digital Image Processing 1

DIGITAL IMAGE PROCESSING


Interest in digital image processing methods stems from two aspects:

1) Improve images for human interpretation;

2) Process images, include storage,transmission and representation for autonomous machine perception.

Why do we need to study DIP?


. An image may be defined a two-dimension function f(x,y)

(x,y) ---- coordinate for a point in a plane.

f ---- intensity or gray or color in the position (x,y)

. Digital image

digitize to the function f and coordinates (x,y) let them

become discrete values.

Usually f,x and y are all finite values.

. Pixel

a point in a digital image is called pixel.

. Digital Image Processing

regarded as a discipline from an image to another.

1.1 What is a digital image ?


1.2 The Origins of Digital image Processing . The first application was in newspaper industry in

1920s


Digital Image Processing, 2nd ed.Digital Image Processing, 2nd ed.www.imageprocessingbook.com

© 2002 R. C. Gonzalez & R. E. Woods

Chapter 1: IntroductionChapter 1: Introduction

They were not real image processing, only image encoding and printing and some improvements.

In 1960s, computer and its programming were brought into, the true image processing began.

Two important events push forward digital image processing.1) Space Program --- First Moon Probe ( America, in 1964);2) In medicine --- CAT or CT (Computer Axial Tomography early 1970s) Using X-rays generates image.


1.Improve qualities of images so that human can interpret

them better.

Such as enhancement, restoration and so on.

2.Process pictures and extract some information from them

for machine perception.

Such as image analysis, image recognition and so on.

1.3 Objectives of digital image processing


.Low-level processes:reduce noises,contrast enhancement and so

on, from an image to another, improve the image quality;

.Mid-level processes: extract some attributes from an image,

segment an image, extract object contour in an image

.High-level processes: recognize objects in an image for analysis

There is no obvious boundary between digital image processing and computer vision.

Computer vision: Machine perception based on vision or to use computers to emulate human vision. Image recognition is a little like this.

1.4 Three level-processes for a digital image


. Digitizing an image ( convert an continuous image to

a digital one)

. Enhancing an image ( Let an image better suit for a specific

application)

. Restoring an image ( Recover a damaged image)

. Compressing an image ( Store it with less bytes )

. Segmenting an image ( Partition objects in an image from

background )

. Recognizing an image ( Tell what the objects are in an image )

1.5 What can digital image processing do?


x

F(x,y)

Y

X

. Sample ( like these grids)

. quantization

1.5.1 Digitizing an image

It is the first step of digital image processing


.

Original Image

Enhanced Image

1.5.2 Enhancing an image


When an image damaged, we can recover it

Cracked parts

torn

1.5.3 Restoring an image


Original image 257kb Compressed image 147kb

Redundant Info.

1.5.4 Compressing an image


Original image Segmented image

Segmenting an image


Take car license plate recognition as an example.

1.5.6 Recognizing an image


. The differences between them can be shown as follows.

image processingImage1 Image2

Computer GraphicsData Image3

Image1 and Image2 are different: Image2 is gotten by processing image1;Image3 is produced or generated by converting data, which maybe a virtual image; Some examples are as follows.

1.6 Digital image processing and computer graphics.


A simple example for computer graphics is that when we input the center coordinates (x,y) and a radius R, a circle ( image) can be produced by computer graphics system.

(x,y)

R

Some examples about their differences.


Camera or Scanner

Original image

Digitization

Pre-processing

Enhancement

Processing

RestorationCompressionRecognitionAnalysis

Processed image

An interpretation

Control Signals

Controlled Devices

1.7 The flow of a typical digital image processing system


1.8 The elements of a digital image processing system

Image acquisition: Digital camera or scanner or video camera;

Image storage: all kinds of digital memory, such as hard disk,

tape, optical disk and so on;

Image processing: Computers with software;

Image display: Displayer or all kinds of hardcopy devices.


It is widely used in industry,medical image, commerce,

entertainment and so on.

(1) Industry monitoring system

e.g. Temperature control ,

automatically adjust

temperature based on

the color in the flame image.

1.9 Some Applications of digital image processing


. Cracking a criminal case with the help of fingerprint

recognition,

. Personal Identity Card,

Store the fingerprint image in the ID card.

. Entry management

--- The conference permission entrance

--- Large park with multi-entrances

Application of Fingerprint System


The key is car license plate recognition based on

image

--- Acquiring the image of a car license plate

Using camera or video camera

--- Enhancement processing

Adjusting the distribution of the gray level in an

image

--- Segmentation

Segmenting letters or digitals in the plate

--- Recognition

Telling what the letters or digitals are

(3) Traffic Management


. It can be widely used to the following aspects

.. Charge automatically on freeway

--- Auto-record the car license plate

--- Distinguish the type of car

--- Recognize the plate

--- Connecting the credit card system automatically

(4) Traffic Control


.. Park management

Automatically record the car license plate and recognize it

and control passing-bar to switch on or off.

.. Monitor the driver with over-speed on freeway

automatically record the car license plate with over-speed

and recognize it.

(5) Traffic Control


An example --- Human face beautifying

(6) Entertainment


(1) The electromagnetic spectrum arranged according to energy per photon.

(2) X-ray and Visual bands of spectrum are the most familiar images in

actual application, such as X-ray in medical inspection and so on.

1.10 Some examples of Using DIP Based on EM spectrum


(3) Gamma-Ray Imaging

Nuclear medicine: Inject a patient with a radioactive isotope that can emit

gamma-ray. It is used in locating sites of bone pathology,such as infection

or tumors.

PET --- Positron Emission Tomograph



(4) X-ray Imaging

Be widely used in Medical diagnostics, Industry, Astronomy and so on.

When X-rays penetrate an objects, there is a different amount of absorption

for different parts in the object , so an image is generated in the film to be

sensitive to X-ray energy.



(5) Imaging in the Ultraviolet Band

Applications of ultraviolet “light” include lithography,industrialinspection, microscopy,lasers,biological imaging,and astronomicalobservations.

i) Ultraviolet light is used in fluorescence

microscopy.ii) Fluorescence microscopy is an excellent method for studying material that can be made to fluorescene.


(6) Imaging in the Visible and Infrared Bands

The visual band of the electromagnetic spectrum is the mostfamiliar in our activities. So their applications are also the widest compared with other bands. And infrared band is often viewedvisual light.

Here we show some applications of the visual light in light Microscopy,astronomy,remote sensing,industry,and law

enforcement.

i) Some examples of images obtained with a light microscope.


These examples range from pharmaceuticals and

micro- inspection to materials

characterization.


iii) Weather observation and prediction also are major applications of multi-spectral imaging from satellites.


(7) Imaging in the Microwave Band

A typical application in the microwave band is radar.


(8) Imaging in the Radio Band

The major applications in the radio band are in medicine andastronomy. In medicine radio waves are used in magneticresonance imaging (MRI).


Comparison with other bands


For an image we must digitize it so that it can be processed by computers.

For an image, we usually use the intensity function f(x,y) to represent it. (x,y) --- the location of a point in the image; f (x,y) --- the intensity of the point (x,y);

It is obvious that 0< f(x,y) <

1.11 How to digitize an image


A simple model is f(x,y)= i(x,y)r(x,y)

i(x,y) --- intensity of the incident light

0 < i (x,y) <

r(x,y) --- the coefficient of the reflection, depend on the object light casts 0 < r(x,y) < 1


Take a picture as an example.

f (x,y)

Y

X


Sampling --- Digitize the spatial coordinates ( pixel )

Quantizing --- Digitize the intensity function f (x,y)

An image processed by sampling and quantizing is called the digital image. It is also the procedure from a continuous image to a discrete one.

Uniform sampling --- If all sampled points are equal spaces Uniform quantizing --- If all grey-level intervals are the same ( From the darkest to the brightest )


Suppose there are N pixels along horizontal direction X and M pixels along vertical direction Y. and there are L gray levels, a digital image cab be represented by the following matrix.

)1,1()1,1()0,1(

....................

....................

)1,1(........)1,1()0,1(

)1,0(........)1,0()0,0(

),(

MNfNfNf

Mfff

Mfff

yxf


For any point (x,y) in the digital image

1,......1,0 Nx 1,......1,0 My

1,......1,0),( Lyxf

Usually, 2n

N 2m

M and 2k

L

So the number, b, of bits required to store a digital image:

kNMb


As for the quality of a digital image, the larger are M, Nand L, the better is the image. For a square image, we have M=N, so

Usually

mb N 2

Image processing

Education

image original image

image recognition

image analysis

damaged image

image damaged

digital image processing7

digital image processing12

digital image processing18