VISVESVARAYA TECHNOLOGICAL UNIVERSITY

Jnana Sangama, Belagavi – 590 014

A Database Management System Mini Project Report On

“TITLE OF THE PROJECT”

Submitted in Partial fulfillment of the Requirements for the V Semester of the Degree of
Bachelor of Engineering
In
Computer Science & Engineering
By

NAME (USN) (16)

NAME (USN) (16)

Under the guidance of

Dr. NAGARAJ BHAT

Associate Professor and Head

Department of Computer Science and Engineering

SHRI MADHWA VADIRAJA INSTITUTE OF TECHNOLOGY AND MANAGEMENT
Vishwothama Nagar, BANTAKAL – 574 115, Udupi District

JANUARY, 2020
SHRI MADHWA VADIRAJA INSTITUTE OF TECHNOLOGY AND MANAGEMENT
(A Unit of Shri Sode Vadiraja Mutt Education Trust ®, Udupi)
Vishwothama Nagar, BANTAKAL – 574 115, Udupi District, Karnataka, INDIA

Department of Computer Science and Engineering

CERTIFICATE

Certified that the Database Management System Project Work titled ‘……………………………’ has

been carried out by Mr. XYZ (4MW10CSXXX) and Ms. ABC (4MW10CSXXX), who are the

bonafide students of Shri Madhwa Vadiraja Institute of Technology and Management, in partial

fulfillment for the award of Bachelor of Engineering in Computer Science and Engineering of

Visvesvaraya Technological University, Belagavi during the year 2020-21. The Database

Management System Mini Project Report has been approved as it satisfies the academic

requirements with respect to the project work guidelines prescribed for the said Degree.

Dr. NAGARAJ BHAT

Project Guide & HOD
Dept. of CSE

External Viva

Name of the Examiners: Signature with Date

1.

2.
 Acknowledgements
________________________________________________

We express our deepest gratitude and respect to our guide Mr. XYZ, Senior Lecturer,
Department of Computer Science and Engineering, for his valuable guidance and
encouragement while doing this project work.

We are indebted to Prof. Dr. Thirumaleshwara Bhat, Principal, for their advice and suggestions at
various stages of the work. We also extend our heartfelt gratitude to Mr. PQR for his assistance.

We extend our thanks to the Management of Shri Madhwa Vadiraja Institute of Technology and
Management, Bantakal, Udupi for providing good laboratory and library facilities. We also
remain grateful to the co-operation and help rendered by the teaching and non-teaching staff of
the Computer Science and Engineering Department.

Lastly, we take this opportunity to offer our regards to all of those who have supported us
directly or indirectly in the successful completion of this project work.

Name of the Student

 ABSTRACT
Remote sensing (RS) in Earth’s perspective is the process of obtaining information about the Earth’s
surface features without being in direct contact with it, but using on-board sensors or camera systems
from the satellite platform [1]. During the past few decades, the majority of remote sensing work has been
focused only on natural environments as the availability of data was restricted to low and medium
resolution images.
 Table of Contents

Page No.

Acknowledgements i

Abstract ii

Table of Contents iv

List of Figures viii

List of Tables xii

List of Acronyms and Abbreviations xvi

Chapter 1 Introduction 1

1.1 Remote sensing 1

1.2 Classification and related constraints 3
1.3 The present study 5

Chapter 2 Literature Review 7

2.1 General 7
2.2 Image fusion 7
2.2.1 Conventional image fusion algorithms 10
2.2.1.1 Multiplicative technique
11
2.2.1.2 Brovey transform 11
2.2.2 Evaluation of image quality 13
2.3 Classification process 15
2.3.1 Sampling schemes 17
2.3.1.1 Sampling unit 17
2.3.1.2 Sampling design 17
2.3.2 Quantitative metrics of class separation 20
2.3.2.1 Transformed divergence
21
2.3.2.2 Jeffries-Matusita 22
2.3.3 Post-classification smoothing 22

iv
 2.4 Classification techniques 26
2.4.1 Unsupervised classification 27
2.4.2 Supervised classification 28
2.4.2.1 Parallelepiped classifier 28
2.4.2.2 Minimum distance-to-means classifier 29
2.4.3 Non-parametric classifier 32
2.4.3.1 Support vector machines 33
2.4.3.2 Artificial neural network 34
2.4.4 Pixel unmixing and Soft classification 36
2.4.5 Object-based classification 37
2.5 Ancillary data 38
2.5.1 Colour space transformation: Intensity-Hue-Saturation 39
2.5.3 Texture 40
2.5.3.1 Approaches to texture analysis 42
2.5.3.2 Texture analysis in remote sensing 43
2.6 Summary 45

Chapter 3 Decision Tree Classification algorithm and GLCM Texture 48

3.1 General 48
3.2 Decision tree classification algorithm 49
3.2.1 Attribute selection criteria 51
3.2.1.1 Criteria for evaluating splits-Information theory 52
3.2.2 Pruning of decision trees 54
3.2.5 Factors affecting the performance of decision trees 56
3.3 GLCM Texture 57
3.3.1 Computation of GLCM 57
3.3.2 Texture measures 60
3.3.3 Parameters affecting texture 61
3.4 Summary 64
Chapter 4 Problem formulation 66

4.1 General 66
4.2 Problem statement 66
4.4 Aim of the work 71
4.5 Objectives of the present study 72
4.6 Summary 74

v
Chapter 5 Materials and Methodology 75

5.1 General 75
5.2 Materials and Data products 76
5.2.1 Data used 76
5.2.2 Instruments used 76
5.3 Study area 77
5.4 Research approach: Phase-I 78
5.4.1 Training and validation datasets 78
5.6 Research approach: Phase-III 86
5.6.1 Texture: Selection of optimal window size
and interpixel distance 86

5.6.2 Investigation of texture at class hierarchy level I and II 88

5.7 Summary 90

Chapter 6 Results and Discussion 91

6.1 General 91
6.2 Result analysis: Phase-I 92
6.2.1 Evaluation of image fusion techniques 92

6.4 Result analysis: Phase-III 136

6.4.1 Texture: Selection of optimal window size and 136
interpixel distance
6.4.2 Effectiveness of texture feature combinations 142

6.5 Summary 177

Chapter 7 Conclusions and Future Enhancements 178

7.1 General 178

7.2 Conclusions 179
7.2.1 Performance study of image fusion algorithms and
evaluation of the reliability of quality assessment measures 179

7.2.2 Performance study and employability of decision

tree classification algorithm on multi-spectral data 180

7.3 Scope of future work 183

7.4 Summary 185

vi
References 189

Appendix-A Figures 205

Appendix-B Tables 212

Personal Profile 213

vii
 List of Figures

Page No.
Figure 2.1 Flowchart of PCA image fusion technique 12

Figure 2.2 Spectral distribution of two classes in a single band data 21

Figure 2.3 2-D scatter plot of parallelepiped classification strategy 28

Figure 2.4 2-D scatter plot of minimum distance-to-means classification strategy 29

8
 List of Tables

Page No.
Table 2.1 Example of an error matrix 23

Table 3.1 Example dataset of a decision tree 51

Table 3.2 Definitions of the GLCM texture measures used in the study 60

Table 5.1 Details of the data products used 76

Table 5.2 LU/ LC class hierarchy levels and details of the training and validation 79
datasets: Phase I and II

Table 5.3 LU/ LC class hierarchy levels and details of the training and validation 88
datasets: Phase-III

Table 6.1 Histogram-means of the bands of the fused images 99

Table 6.2 Entropy of the bands of the fused images 99

9
 Chapter 1

Introduction

1.1 Remote sensing

R
emote sensing (RS) in Earth’s perspective is the process of obtaining information about the
Earth’s surface features without being in direct contact with it, but using on-board sensors or
camera systems from the satellite platform [1]. During the past few decades, the majority of
remote sensing work has been focused only on natural environments as the availability of data
was restricted to low and medium resolution images [2].

Unless a critical study of the above mentioned criteria or indices is made in respect of reliability
and relevancy to the context, their non-rational way of adoption would adversely affect the
performance of even the best classifier. Hence, improving classification accuracy of the remote
sensing data has always been an important concern to extract the real world situation in the
form of thematic maps. A similar conclusion is also drawn in literature [3], [4], [5].

10
 Chapter 4 (16)

Problem formulation
(Chapter Heading 18)

4.1 General (Side Heading Font 16)

Chapter 4 outlines the present-day-challenges in RS classification, motivation to the present
study, problem statement and research objectives of the work. The literature survey discussed in
chapter 2 identifies image fusion, classification algorithms and effective integration of ancillary
data into classification as the three important areas need to be explored to enhance the
classification accuracy.

4.2 Problem statement (Side Heading Font 16)

4.2.1 First method (Sub Heading Font 14)
From sections 2.2, 2.4, 2.5 and [6] it is intuitive in a broader way that obtaining satisfactory
classification accuracy over urban or semi-urban LU/ LC classes, particularly in high spatial
resolution images, is deemed to be a present day challenge in remote sensing and is attributed
to the following reasons. For classification of features in urban area, the spatial resolution of the
data should be at least 5m or less. [26].

4.2.2 Second method (Sub Heading Font 14)

From sections 2.2, 2.4, 2.5 and [6] it is intuitive in a broader way that obtaining satisfactory
classification accuracy over urban or semi-urban LU/ LC classes, particularly in high spatial
resolution images, is deemed to be a present day challenge in remote sensing and is attributed
to the following reasons. For classification of features in urban area, the spatial resolution of the
data should be at least 5m or less. [26].

4.3 Summary (Side Heading Font 16)

This chapter reviewed the philosophy underlying image fusion, classification procedures and
ancillary data, more importantly the texture, employed in remote sensing. Each section
discussed some associated studies corresponding to the topics to highlight the potential
strengths and weaknesses of each technique.

11
To sum up, the review of literature clearly indicates that the finer spatial resolution data poses
challenge in generating a satisfactory classified image over urban/ semi-urban areas using
traditional classifiers, and the task is not as straightforward as classification of low resolution
imagery (30m or more).

In general, it is apparent from the literature that today, in remote sensing, there is a great need
for an alternative but efficient classification approach which include both advanced classification
algorithms and image processing techniques to exploit the potential information available
through advanced sensor systems and ancillary data bands.

12
 References

[1] Nagaraj Bhat, K C Gouda, Dines Dwivedi, Anurag Patro (2014) Migration Management in
Cloud Computing, International Journal of Engineering Trends and Technology (IJETT),
Vol 12, number 8, ISSN: 2231-5381.
(For research/ technical papers)

[2] M. Paul Mather, Computer Processing of Remotely Sensed Data: An Introduction, John
Wiley and Sons, Ltd., 3rd Edition, 1999.
(For text books)

13