vii
TABLE OF CONTENTS
CHAPTER
TITLE
DECLARATION
DEDICATION
ACKNOWLEDGEMENT
ABSTRACT
ABSTRAK
TABLE OF CONTENTS
LIST OF TABLES
LIST OF FIGURES
LIST OF ABBREVIATIONS
LIST OF SYMBOLS
LIST OF APPENDICES
1
INTRODUCTION
1.1
Background of the Study
1.2
Statement of the Problem
1.3
Objective of the Study
1.4
Scope of Work
1.5
Methodology
1.6
Report Outline
2
LITERATURE REVIEW
2.1
Smart Camera
2.1.1
Classification of Smart Cameras
2.1.2
Functionalities of ASIP
2.2
Canny Algorithm
2.2.1
Image Smoothing
2.2.2
Edge Detection
2.2.3
Directional Non-maximum Suppression
(DNS)
PAGE
ii
iii
iv
v
vi
vii
x
xii
xiv
xv
xvi
1
1
1
2
2
2
3
4
4
5
7
7
9
10
14
viii
2.3
2.4
2.5
2.6
3
2.2.4
Thresholding
Altera Cyclone II Board
2.3.1
DE2 Control Panel Facility
NTSC and PAL Standards
Bitmap Image
Motivation of Extended Work
15
16
17
17
18
18
MODIFIED CANNY ALGORITHM
3.1
Implementing RGB to GrayScale
3.2
Implementing Smoothing in Hardware
3.3
Computing Magnitude and Orientation of Gradient
of Image
3.4
Implementing Thinning in Hardware
3.5
Implementing Hysteresis
3.6
Chapter Summary
22
22
23
4
HARDWARE ARCHITECTURE
4.1
Chapter Summary
31
34
5
RESULTS AND DISCUSSION
5.1
Matlab Results
5.1.1
RGB to GrayScale
5.1.2
Smoothing
5.1.3
Magnitude and Direction of Gradient
5.1.4
Directional Non-maximum Suppression
(DNS)
5.1.5
Hysteresis
5.2
Co-Design Results
5.2.1
All Steps in NiosII
5.2.2
Computing only Grayscale in Hardware
and the Rest in NiosII
5.2.3
Computing only Smoothing in Hardware
and the Rest in NiosII
5.2.4
Computing only Magnitude and Direction
of Gradient and the Rest in NiosII
5.2.5
Only Thinning in Hardware and the Rest
in NiosII
35
35
35
35
38
25
29
29
30
38
39
42
42
42
43
45
46
ix
6
CONCLUSION
6.1
Future Works
REFERENCES
52
53
54
x
LIST OF TABLES
TABLE NO.
2.1
2.2
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
5.9
5.10
5.11
5.12
5.13
5.14
TITLE
Types of smart cameras, typical characteristics, and
sample applications.
Addresses of different information located in a bitmap
header file.
Execution time for every step of the canny algorithm in
software.
Execution time when only computing grayscale is in
hardware.
Numbers of SOPC builder generated module resources
when only computing grayscale is in hardware.
Numbers of computing grayscale designed module
resources.
Execution time when only smoothing in hardware using
method1.
Execution time when only smoothing in hardware using
method2.
Numbers of SOPC builder generated module resources
when only smoothing is in hardware.
Numbers of smoothing designed module resources.
Execution time when only computing gradient is in
hardware.
Numbers of SOPC builder generated module resources
when only computing gradient is in hardware.
Numbers of computing gradient designed module
resources.
Execution time when only thinning is in hardware.
Numbers of SOPC builder generated module resources
when only thinning is in hardware.
Numbers of thinning designed module resources.
PAGE
5
20
42
43
43
43
44
44
45
45
45
46
46
46
47
47
xi
5.15
5.16
5.17
5.18
5.19
5.20
Execution time for every step of the canny algorithm
implemented in hardware.
Execution time difference between hardware and
software of the canny algorithm.
Execution time when computing grayscale, smoothing
and computing gradient are in hardware.
Numbers of SOPC builder generated module resources
for final co-design.
Numbers of final co-design, designed module resources.
Execution time for different operations.
47
48
48
48
49
50
xii
LIST OF FIGURES
FIGURE NO.
1.1
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
2.10
2.11
2.12
2.13
3.1
3.2
3.3
3.4
3.5
3.6
3.7
TITLE
Steps of implementing algorithms on hardware.
Functional structure of a smart camera.
Classification of smart cameras based on levels of
integration.
Canny algorithm steps.
Gaussian shape in respect to different standard deviations.
Gaussian kernel.
Image results of a Gaussian function with different
standard derivations applied to an image.
Additive Gaussian noise applied to images of rows
number 2, 3 and 4 with a standard deviation of 0.1,1 and
10 , respectively.
Roberts, Prewitt and Sobel masks using for implementing first order derivative in x and y direction.
An example of how gradient works.
An example of how thresholding works.
Altera DE2 development and education board.
The DE2 control panel concept.
Bitmap header file.
Gaussian function.
Graphical show of convolution an image with Gaussian
mask.
Sobel masks in x and y directions.
Central pixel and directions of it’s neighbors.
Division of trigonometric circle to four directions.
Upper half of trigonometric circle.
Black central pixel and its corresponding neighbors for
thinning, based on the direction.
PAGE
3
4
6
9
10
10
10
11
13
14
15
16
17
19
24
24
25
26
27
28
29
xiii
3.8
4.1
4.2
4.3
4.4
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
5.9
5.10
5.11
Black central pixel and its corresponding neighbors for
thresholding, based on the direction.
Sequence of canny algorithm implementation.
Hardware architecture of final design.
Sequence of passing data from one step to another one.
Sequence of input/output data to/from system.
Grayscale results using three different methods.
Smoothed results using Gaussian mask.
The measured bit difference between smoothed images
using original and estimated Gaussian mask.
Magnitude of gradient.
Simple thresholded and thinned image results
(T h=100).
A horizontal edge with 3 different values.
Image results before and after hysteresis (T h=200 and
T l=50).
Image results before and after hysteresis (T h=100 and
T l=50).
Matlab results of modified Canny algorithm.
A set of pixels; central pixels and corresponding
neighbors that should be send to hardware for
smoothing.
24bit input image versus binary output image.
30
31
32
33
34
36
37
37
38
39
39
40
41
41
44
49
xiv
LIST OF ABBREVIATIONS
ASIC
-
Application Specific Integration Circuit
ASIP
-
Application Specific Information Processing
ATSC
-
Advanced Television Systems Committee
DAC
-
Digital to Analog Converter
DNS
-
Directional Non-maximum Suppression
DSP
-
Digital Signal Processor
FPGA
-
Field Programmable Gate Array
HD
-
High Definition
IP
-
Internet Protocol
LED
-
Light Emitting Diode
NTSC
-
National Television Standards Committee
PAL
-
Phase Alternating Line
PLD
-
Programmable Logic Device
RTL
-
Register Transfer Level
SDRAM
-
Synchronous Dynamic Random Access Memory
SOPC
-
System on a Programmable Chip
SRAM
-
Static Random Access Memory
USB
-
Universal Serial Bus
VTR
-
Video Tape Recorder
xv
LIST OF SYMBOLS
Tl
–
Low Threshold Value
Th
–
High Threshold Value
Hz
–
Hertz
M Bps
–
Megabytes Per Second
M bps
–
Megabits Per Second
M
–
Mega
σ
–
Standard Derivative
–