ABSTRACT
This project research aimed to develop a standalone computerapplicationfor image classification using Matlab 9.2 by adopting the Rapid Application Development Model (RAD). The software can be used for image preprocessing and image classification. The test run of the software was carried out using acquired LANDSAT image and Google earth image of Federal University of Technology Gidan-Kwano campus. The program was designed to accept the input image in all format for the classification and automatically classified the image by chosen the appropriate method. The outcome of the software was compared with result gotten from the commercial software (ERDAS) and the statistical analysis of the processing time taking for the execution by the two software were affirmed by IBM SPSS software using two ways ANOVA. The outcome of the analysis confirmed that there is no significant different between the classified image by two software using the same method and there is significant different between there time of execution. The ZANNYCLASS 1.0 was able to come to existence by this project research and for efficient and effectiveness of the software its functionality need to be upgraded by further studies.
CONTENTS
COVER PAGE
TITLE PAGE
DECLARATION i
CERTIFICATION ii
DEDICATION iii
AKNOWLEDGEMENT iv
ABSTRACT v
TABLE OF CONTENTS vi
LIST OF TABLES xii
LIST OF FIGURES xiii
CHAPTER ONE 1
INTRODUCTION 1
1.0 Background to the Study 1
1.1 Statement of Problem 3
1.2 Aim and Objectives of the Study 4
1.2.1 Aim 4
1.2.2 Objective 4
1.2 Scope of the Project 4
1.4 Study Area 5
1.5 Limitation 5
CHAPTER TWO 6
LITERATURE REVIEW 6
2.0 Theoretical Concept 6
2.1 Supervised 6
2.1.1 Minimum Distance Algorithm 6
2.1.2 K-Nearest Neighbor Algorithm 7
2.1.3 Nearest Clustering Algorithm 8
2.1.4 Fuzzy C-Means Algorithm 8
2.1.5 Maximum Likelihood Algorithm 9
2.1.6 Artificial Neural Network Algorithm (ANNs) 10
2.2 Unsupervised 12
2.2.1 K-Means Clustering Algorithm 12
2.2.2 Isodata 13
2.3 Software Development Approaches 13
2.3.1 Rapid Application Development Model 14
2.3.2 Spiral Model 15
2.3.3 Iterative Model 16
2.3.4 Prototype Model 17
2.3.5 V- Model 18
2.3.6 Water Fall Model 20
2.3.6 Evolutionary Software Development Model 21
2.4 Platform Chosen 22
2.5 Image Classification Softwares 23
2.5.1 Commercial Softwares 23
2.5.1.1 Idrisi 23
2.5.1.2 Erdas 24
2.6 Gap to Be Filled 26
CHAPTER THREE 27
METHODOLOGY 27
3.0 General overview 27
3.1 Requirement Planning 28
3.1.1 System Definition 28
3.1.2 Hardware and Software Used 29
3.1.3 Data Source 30
3.2 Requirement Definition 30
3.2.1 Unsupervised Classification 30
3.2.1.1 Clustering Stage 31
3.3 User Design Phase 32
3.3.1 Flow Chart: 32
3.3.1.1 Subroutine for software adopted process 33
3.3.1.2 Subroutine for K-means method 34
3.3.2 Program Layout Design 35
3.3.3 Designation of Graphical User Interface 35
3.4 Construction Stage 38
3.5 Implementation of System 38
CHAPTER FOUR 40
RESULTS AND DISCUSSION 40
4.1 Result Presentation 40
4.1.1 Result of Unsupervised Classification Using Landsat Image 40
4.1.1.1 Time Consideration and Processing Speed 41
4.1.2 Result of Unsupervised Classification Using Google Earth Image 43
4.1.2.1 Time Consideration and Processing Speed 44
4.1.3 Comparison of the Outputs with Other Software 45
4.1.3.1 Time Consideration and Processing Speed 46
4.1.4 Result of Unsupervised Classification Using Google Earth Image 48
4.2 Result Analysis 52
4.2.1 The Design 52
4.2.2 The Data 52
4.2.3 Hypothesis Testing 52
4.2.4 Statistical Test for the Chosen Hypothesis 53
4.2.5 Discussion of Results 53
CHAPTER FIVE 55
SUMMARY, CONCLUSION AND RECOMMENDATIONS 55
5.1 Summary 55
5.2 Conclusion 56
5.3 Recommendations 56
References 57
Appendix (Codes) 61
