INTELLIGENT VISION SYSTEM FOR DYNAMIC 
ENVIRONMENTS 
 
 
 
A thesis submitted to the 
Department of Electrical Engineering, University ofMoratuwa 
in partial fulfillment of the requirements for the 
Degree of Master of Engineering 
 
 
 
By 
H.Y. ARUNA HEWAWASAM 
 
 
 
Supervised by: Dr. Lanka Udawatta 
 
 
Department of Electrical Engineering 
University of Moratuwa, Sri Lanka 
 
2005 
 
83813 
 
 
 
  
Abstract  
  
This dissertation describes an intelligent vision system that absorbs useful 
information from its environment and draws useful conclusions. This system can 
give the instructions to locate vacant seats that are currently occupying in a cinema 
theater. Extraction of useful information without viewing or exposing inside details 
of an environment through an active vision system is proposed. Reasoning based 
conclusions are drawn for optimum searching. The effectiveness of the proposed 
method is demonstrated using an experiment. 
 
Three reasoning criteria are developed and experimentally tested for identifying the 
states of seats, States of seat can be vacant state, occupied state, or a state with an 
object placed on the seat. First criterion basically uses binary image analysis and 
with the introduction of white reference value it can also be applied for environments 
where there are intermittent variations of illumination level. Second criterion is based 
on the analysis of color image and it can be basically used for identifying objects 
placed on seats. Third criterion based on the analysis of intensity image. 
 
Intelligent vision system was developed using the combination of first and second 
criteria. The created graphical user interface provides links for setting up the system, 
and setup program i provides an interface and instructions for user to find seat 
locations and entering those locations in the main program and other setup programs. 
Setup program 2 is given for automatically calculating the other necessary 
parameters and white reference program for setting up white reference values. 
 
The intelligent vision system can be further developed and generalized for other 
applications.Mainly it can be used for intelligent building applications. For example 
in designing an intelligent room where the movements and changes occurring inside 
the room could be monitored using a camera system. In a multi storey building, 
required information of a particular floor that is used for common seating could be 
displayed at other floors. In a vehicle park, the registration number and the entering 
  
time of the vehicles could be recorded. Available parking spaces can be displayed at 
the entrance. 
DEC LARA TJON 
The work submitted in this dissertation is the result of my own investigation, except 
where otheiWise stated. 
It has not already been accepted for any degree, and is also not being concurrently 
submitted for any other degree 
~± -
H.Y.Aruna Hewawasam 
July 20, 2005 
I endorse the declaration by the candidate. 
___ [1£1!E__ --········--
Dr. Lanka Udawatta 
j 
-.. 
ACKNOWLEDGMENTS 
Thanks are due first to my supervisor, Dr. Lanka Udawattha, for his great insights, 
perspectives, guidance and sense of humor. My sincere thanks should also go to the 
other lectures, Prof. Lucas, Prof. Ranj ith Perera, and Prof. Sriyananda, who gave 
instructions and pointed out shortcomings during my presentations. 
l 
Sincere gratitude is also extended to the people who serve in the Department of 
Electrical Engineering, University of Moratuwa, Sri Lanka for helping in various 
ways to clarify the things related to my academic works in time with excellent 
cooperation and guidance. 
I should not forget the corporation and the support given by my family members, my 
wife, parents, and brothers. May be, I could not have made it without their support. 
I would also like the thank all of my friends who supported me in this attempt 
specially helping me to get pictures and setting up the camera etc 
Lastly, I should thank many individuals, friends and colleagues who have not been 
mentioned here personally in making this educational process a success . 
._. 
H. Y.Aruna Hewawasam 
July 20, 2005 
LIST OF FIGURES 
Figure Page 
1.1 Intelligent Vision System Architecture 
1.2 RGB Image Structure 5 
2.1 Use of Marks on Seats 16 
2.2 System Architecture for the Experiment 17 
3.1 Repeated Dilations of Marker Image, Constrained by Mask/ 24 
3.2 Intermediate Stages in Processing a Color Image 25 
3.3 Matrices Ak and Ak 26 
3.4 Use of Color Image Matrix 27 
3.5 Creating Color Components Matrices 28 
3.6 Improving the accuracy 33 
(a) Improved seat marks 
(b) Fine thick edge marks 
3.7 Effect of Enhancement 36 
4.1 Seat Arrangement 37 
4.2 Hall Arrangement 39 
(a) Effective Length 
(b) Effective Width 
6.1 Achieving Illumination Level Variation 53 
6.2 Illumination Level Variation 54 
6.3 Variation in Color Components with Illumination Level 
for White Color Mark 55 
6.4 Variation in Color Components with mumination Level 
--
for Blue Color Mark 55 
6.5 Variation in Color Components with Illumination Level 
for Yell ow Color Mark 56 
6.6 Testing Different Types oflmages 58 
6.7 Testing Different Types of Images 59 
6.8 Testing Different Types of Images 60 
6.9 Object Identification- Result 1 61 
6.10 Object Identification - Result 2 62 
7.1 Graphical User interface 64 
7.2 Graphical User Interface Showing Setup Menu Options 65 
7.3 Graphical User Interface Showing Help Menu Options 65 
7.4 Results of Setting up Program 1 l 66 
... 
7.5 Graphical User Interface Showing a Stage of Demo 1 68 
7.6 Results of Demonstration 3- Result 1 69 
7.7 Results of Demonstration 3 - Result 2 69 
--
LIST OF TABLES 
Table 
3.1 Condition Table 
4.1 Example for Determination of 'd' Values 
6.1 Variation in Color Components with Illumination Level 
6.2 Summary of Object Identification Result 
i 
-" 
----
Page 
34 
42 
56 
63 
CONTENTS 
Declaration 
Abstract 
Acknowledgement 
List of Figures 
List of Tables 
Chapters 
1 
2 
Introduction 
I .I Background and Literature Survey 
1.2 ~otivation 
1.3 Goals 
1.4 Achievement m brief 
System Overview 
2.1 The approach 
2.1.1 White reference 
2.1.2 Algorithms 
2.2 System components 
2.2.1 Camera 
2.2.2 Transmitter receiver system 
2.2.3 Computer 
2.2.4 Lights 
I 
-" 
---
Page 
i 
ii 
iii 
IV 
vii 
01 
02 
13 
14 
15 
16 
16 
17 
18 
19 
19 
20 
20 
20 
3 Image analysis 22 
3.1 First reasoning criterion 22 
3.2 Second reasoning criterion 27 
3 .2.1 Determination of mark locations 27 
3.2.2 Analysis for mark 1 28 
3 .2.3 Analysis for mark 2 30 
3.2.4 Analysis for mark 3 31 
3.2.5 Ultimate reasoning 32 
3.3 Third reasoning criterion 32 
3.4 How to use criteria 1, 2 & 3 33 
3.4.1 Condition table j 34 
3.5 Detection of illumination level variation 35 
3.5.1 White reference 35 
3.6 Image enhancement 35 
4 Setting up the system 37 
4.1 Setup program 1 39 
4.2 Setup program 2 41 
4.3 White reference program 43 
5 Image acquisition and processing 45 
5.1 Image acquisition 45 
5.1.1 Basic image acquisition procedure 45 
5.1.2 Image acquisition algorithm 49 
5.2 Image processing 49 
5.2.1 Running algorithm 50 
Experimental Results 
..... 53 6 
6.1 Intensity variation 53 
6.2 Testing for different conditions 57 
6.3 Object identification 60 
6.3.1 Summary of the results 63 
7 
8 
Graphical User Interface 
7.1 Setup program 1 
7.2 Setup program 2 
7.3 White reference program 
7.4 Demonstration 1 
7.5 Demonstration 2 
7.6 Demonstration 3 
Concluding Remarks and Further Developments 
8.1 Conclusions 
8.2 Recommendations fo r Future Research 
.f 
·"' 
References 
Appendices 
Appendix A : Graphical user interface program 
Appendix B : Running program 
Appendix C : Setting up program 1 
Appendix D : Setting up program 2 
Appendix E : White reference program 
Appendix F : Demonstration 1 
Appendix G : Demonstration 2 
Appendix H : Demonstration 3 
---
64 
66 
67 
67 
67 
68 
68 
70 
70 
71 
72 
75 
75 
77 
79 
80 
81 
82 
83 
84