Digital Image Processing

Introduction

Image Processing :

Processing

Input Output

Image Image

Examples

Contrast Stretching

1.

Low Contrast Output

Image Image

Noise Filtering

2.

Noisy Noise-free

Image Image

3.

De-blurring/ Deconvolution

Defocused or Focused /

Blurred and Sharp

Noisy Image Image

Image

Compression

4.

Image with Compressed

redundant Image

data

etc.

Applications

1.Editing Digital photos

Image Enhancements

Contrast Stretching

Color Correction

White-Balancing

Red-Eye Reduction/ Removal

Shrinking/Zooming

Software :

Adobe Photoshop

ImageMagick

GIMP

2. Medical Image Processing

Chest X-rays

Computed Tomography

UltraSound Imaging

MRI, PET, etc.

3. Machine Vision

Inspection

Measurement

Robotic Vision

4. Remote Sensing

Land use

Crops

Weather Forecasting

5.Defense

Target recognition

Thermal Imaging

6. Image Storage & Transmission

JPEG,MPEG, etc.

7. OCR, Document Image Processing

Contents

Introduction

Image Fundamentals

Enhancements-Spatial Domain

Enhancements-Freq. Domain

Image Restoration

Color Image Processing

Wavelets & Multi-resolution processing

Image Compression

Image Reconstruction

Binary Image Processing

Image Sensors/Input Devices

Film Cameras

Digital Cameras (CCD/CMOS)

Video Cameras

Scanners

Image Output Devices

Printers

TV/Computer Monitors

Image Storage :

CD-ROM

Hard disk

Processing Hardware :

PCs, Microprocessors, boards, workstations, etc.

Processing Software :

Many commercial and non-commercial

Chapter 2

Digital Image Fundamentals

Image Acquisition by a Digital Camera

Light

Illumination

Source

LCD Array

Scene

Radiance

3D

Scene

Radiance

Lens

Digital Camera

Photometric Relation :

Image Irradiance a Scene Radiance

Geometric Relation :

Perspective Projection

x = (X / Z) * f

y = (Y / Z) * f

FUNDAMENTALS Pixel gray level

Rows Columns

0 1 2 3

1

2

3

4

5

6

7

CCD Array

Photon

		A/D Quantizer

Voltage Digital Output

Charge

Voltage a Image Brightness

A black / white Digital Image is a 2D array of pixels with gray values.

8 bits / pixel – typical

N rows * M columns * 8 bits / pixel bits

Colour Digital Camera

Simple Model

Blue CCD

Green

CCD

Red

CCD

Beam Splitter

Prism

Blue

Green

Red band

Image

M * N

3 digital images

Color Cube

Colour

(R,G,B) = (40,70,90)

G

B

O R

Human Visual System (HVS)

HVS is often the end consumer of the output of image processing algorithms.

Therefore, a knowledge of the HVS characteristics is useful in optimizing the digital image processing algorithms to yield better looking images.

E.g. Image compression algorithms discard high frequency information for which the HVS has lower sensitivity.

3-types

C2

C3

C1

400 nm Green 700 nm

Violet Red

Perceive

f (x , y)

Discrete Sampling

Quantization

Digital Image

Zoom in/out

0 1 2 …………………M-1

0

0 1 2….. N-1

1 0

1

2

.

2

. .

N-1

M-1

X/ (M-1) = x / (N-1)

X = (M-1) / (N-1) x

f(x,y)

for (x=0 ; x< N ; x++)

{

for (y=0 ; y < N; y++)

{

X= (M-1)\ (N-1) x

Y= (M-1)\ (N-1) y

Z = bilinear interpolation

f’(x,y) =Z

}

}