BORANG PENGESAHAN STATUS TESIS* LAND SURFACE TEMPERATURE DETERMINATION FOR

Transcription

1 2 BORANG PENGESAHAN STATUS TESIS* LAND SURFACE TEMPERATURE DETERMINATION FOR JUDUL : SELANGOR AREA SESI PENGAJIAN : 2011 _/ 2012 _ Saya UMMI KALSOM BINTI MOHAMAD HASHIM mengaku membenarkan tesis Projek Sarjana Muda ini disimpan di Perpustakaan Fakulti Teknologi Maklumat dan Komunikasi dengan syarat-syarat kegunaan seperti berikut: 1. Tesis dan projek adalah hakmilik Universiti Teknikal Malaysia Melaka. 2. Perpustakaan Fakulti Teknologi Maklumat dan Komunikasi dibenarkan membuat salinan untuk tujuan pengajian sahaja. 3. Perpustakaan Fakulti Teknologi Maklumat dan Komunikasi dibenarkan membuat salinan tesis ini sebagai bahan pertukaran antara institusi pengajian tinggi. 4. ** Sila tandakan (/) SULIT TERHAD (Mengandungi maklumat yang berdarjah keselamatan atau kepentingan Malaysia seperti yang termaktub di dalam AKTA RAHSIA RASMI 1972) (Mengandungi maklumat TERHAD yang telah ditentukan oleh organisasi/badan di mana penyelidikan dijalankan) TIDAK TERHAD (TANDATANGAN PENULIS) (TANDATANGAN PENYELIA) Alamat tetap: KG.PULAU KABONG DR. ASMALA BIN AHMAD 32700, BERUAS Nama Penyelia PERAK Tarikh: Tarikh: CATATAN: * Tesis dimaksudkan sebagai Laporan Projek Sarjana Muda (PSM). ** Jika tesis ini SULIT atau atau TERHAD, sila lampirkan surat daripada pihak berkuasa.

2 LAND SURFACE TEMPERATURE DETERMINATION FOR SELANGOR AREA UMMI KALSOM BINTI MOHAMAD HASHIM UNIVERSITI TEKNIKAL MALAYSIA MELAKA

3 LAND SURFACE TEMPERATURE DETERMINATION FOR SELANGOR AREA UMMI KALSOM BINTI MOHAMAD HASHIM This report is submitted in partial fulfillment of the requirements for the Bachelor of Computer Science (Artificial Intelligence) FACULTY OF INFORMATION AND COMMUNICATION TECHNOLOGY UNIVERSITI TEKNIKAL MALAYSIA MELAKA 2013

4 ii DECLARATION I hereby declare that this project report entitled LAND SURFACE TEMPERATURE DETERMINATION FOR SELANGOR AREA is written by me and is my own effort and that no part has been plagiarized without citations. STUDENT : Date: UMMI KALSOM BINTI MOHAMAD HASHIM SUPERVISOR : Date: DR. ASMALA BIN AHMAD

5 iii DEDICATION I dedicate my final year project report to my family and many friends. A special to my loving parents, whose words of encouragement and push for tenacity ring in my ears. I also dedicate this report to my close friends and family who have supported me throughout the project development. I will always appreciate all they have done. I dedicate this work and give special thanks to my best friends Fatehah and Amalina for being support me throughout the entire project development. Both of you have been my best cheerleaders.

6 iv ACKNOWLEDGEMENTS I would like to acknowledge the contributions of the following individuals to the development of this final year project report: My class peer for the cooperation and camaraderie. I am also heartily thankful to my final year project supervisor, Dr.Asmala bin Ahmad, whose encouragement, guidance and support from the initial to the final level enabled me to develop an understanding of the project. I would also like to thank Dr.Zuraida and Mr.Ahmad Fadzli Nizam for their comments and questions were very beneficial in my completion of my final year project. To my truly great friends Fatehah Hanim Ahmad Othman and Nuramalina Mat Serah who has made available their support in a number of ways. Lastly, I am offer my regards and blessings to all of those who supported me in any respect during the completion of the project.

7 v ABSTRACT The study was implemented by using Landsat TM. This scene covers the areas of Klang, Selangor. The main objective of this study is to evaluate the use of remote sensed information. Remote sensing is a technology to obtain information about the object or phenomenon without having to be physically connected to the object. Remote sensed information is use to gain land surface temperature (LST) using supervised method of Landsat images. The derivation of LST map using remote sensing technique in this study is useful in providing information for analyzing geophysical information over Selangor area, especially dealing with the urban heat island phenomenon. Supervised image classification was used, in which it requires training sites on the image that represent of each desired land temperature category. The delineation of training areas that represent the land temperature is most effective when an image analyst has knowledge of the geography of a region and experience with the spectral properties of the suitable classes. Envi software is used to identified spectral values or signatures associated with the training sites. After the signatures for each land temperature category have been identified, the software then uses all the signatures to classify the remaining pixels. The methodology consist 4 phases; analysis, design, implementation and testing.

8 vi ABSTRAK Kajian ini dilaksanakan menggunakan Landsat TM. Kejadian ini meliputi kawasan Klang, Selangor. Objektif utama kajian ini adalah untuk menilai penggunaan maklumat Remote Sensing. Remote Sensing adalah perolehan maklumat tentang objek atau fenomena tanpa membuat hubungan fizikal dengan objek. Maklumat remote sensing adalah digunakan untuk mendapatkan suhu permukaan tanah (LST) menggunakan kaedah supervised dan imej satelit. Penghasilan peta LST menggunakan teknik remote sensing dalam kajian ini amat berguna dalam menyediakan maklumat untuk menganalisis parameter geofizik di kawasan Selangor, terutamanya berurusan dengan fenomena haba pulau bandar. Supervised image classification adalah satu kaedah di mana penganalisis mentakrifkan kawasan kecil, yang dikenali sebagai tapak latihan, pada imej yang mewakili setiap kategori suhu tanah yang dikehendaki. Penggarisan wakil kawasan latihan suhu tanah adalah paling berkesan apabila seorang penganalisis imej mempunyai pengetahuan geografi rantau dan pengalaman dengan ciri-ciri spektrum kelas perlindungan. Penganalisis imej kemudian menggunakan perisian untuk mengenali nilai spektrum atau petanda yang berkaitan dengan tapak latihan. Selepas petanda bagi setiap kategori suhu tanah telah ditentukan, perisian kemudian menggunakan petanda untuk mengklasifikasikan baki piksel. Kaedah ini terdiri 4 fasa iaitu analisis, reka bentuk, pelaksanaan dan ujian.

9 vii TABLE OF CONTENTS DECLARATION... ii DEDICATION... iii ACKNOWLEDGEMENTS... iv ABSTRACT... v ABSTRAK... vi TABLE OF CONTENTS... vii-x LIST OF TABLES... xi-xii LIST OF FIGURES... xiii-xiv LIST OF ABBREVIATIONS... xv CHAPTER 1: INTRODUCTION Project Background Problem Statements Objectives Scopes Software Scope Area of Study Scope User Scope Project Significance Expected Output Conclusion CHAPTER 2: LITERATURE REVIEW AND PROJECT METHODOLOGY Introduction LST and remote sensing interest

10 viii 2.3 Supervised Method Maximum Likelihood Neural Network Project Methodology Project Schedule and Milestone Conclusion CHAPTER 3: ANALYSIS Introduction Problem Analysis Requirement Analysis Data Requirement Software and Hardware Requirement Ancillary Data Requirement Conclusion CHAPTER 4: DESIGN Introduction LST process Input Geometric correction Subset/Resize Ocean Masking Cloud Masking Classification using Maximum Likelihood and Neural Network LST modeling LST derivation LST segmentation Output... 46

11 ix Accuracy assessment/testing Conclusion CHAPTER 5: IMPLEMENTATION Introduction Data preprocessing Input data process Geometric correction process Subset/resize process Ocean masking process Cloud masking process Classification using Maximum Likelihood and Neural Network LST modeling process Conversion from Digital Number to Radiance Conversion from Radiance to Brightness Temperature LST derivation LST segmentation Output Accuracy Assessment/Testing Conclusion CHAPTER 6: TESTING Introduction Maximum Likelihood Classification Neural Network Classification Comparison Between Region of Interest (ROI) and Maximum Likelihood Classification Comparison Between Region of Interest(ROI) and Neural Network Classification

12 x 6.6 Comparison Between Maximum Likelihood and Neural Network Classification Comparison Between Maximum Likelihood Map and Temperature Map Conclusion CHAPTER 7: APPLICATION OF LAND SURFACE TEMPERATURE DETERMINATION USING ANDROID Introduction Requirement Sofware Hardware Graphical User Interface Codes XML Codes Java Codes Conclusion CHAPTER 8: CONCLUSION Introduction Strengths Weakness Proposition for Improvement Conclusion REFERENCES... xvi-xvii APPENDICES... xviii-xx

13 xi LIST OF TABLES TABLE TITLE PAGE Table 1.0 List of software required 4 Spectral range of bands and spatial resolution for the Table 1.1 Landsat 5 TM sensor 6 Table 2.0 LandSat TM spectral consideration 11 Table 2.1 FYP schedule and milestone 19 Spectral range of bands and spatial resolution for the Table 3.0 Landsat 5 TM sensor 26 Table 3.1 Envi 4.5 Specifications 29 Table 3.2 Mathlab R2010a Specifications 31 Table 3.3 List of Hardware Requirement 31 Table 5.0 Sources of Geometric Error 49 Table 5.1 Image value before and after subset 55 Table 5.2 List of pixel changes at different temperatures 71 Table 5.3 Total of pixel at different temperatures 71 Confusion Matrix in Pixels; Region of Interest vs Table 6.0 Maximum Likelihood 79 Confusion Matrix in Percent; Region of Interest vs Table 6.1 Maximum Likelihood 79 Confusion Matrix in Commission and Ommission; Region Table 6.2 of Interest vs Maximum Likelihood 80 Confusion Matrix in Prod. Acc and User Acc; Region of Table 6.3 Interest vs Maximum Likelihood 80 Confusion Matrix in Pixels; Region of Interest vs Neural Table 6.4 Network 83 Confusion Matrix in Percent; Region of Interest vs Neural Table 6.5 Network 83 Confusion Matrix in Commission and Ommission; Region Table 6.6 of Interest vs Neural Network 84

14 Confusion Matrix in Prod. Acc and User Acc; Region of Table 6.7 Interest vs Neural Network 84 Confusion Matrix in Pixels; Maximum Likelihood vs Table 6.8 Neural Network 87 Confusion Matrix in Percent; Maximum Likelihood vs Table 6.9 Neural Network 87 Confusion Matrix in Commission and Ommission; Table 6.11 Maximum Likelihood vs Neural Network 88 Confusion Matrix in Prod. Acc and User Acc; Maximum Table 6.12 Likelihood vs Neural Network 88 Table 7.0 Specification of laptop 94 Table 7.1 Specifications of tablet 95 Table 8.0 Result of Classification Error Between ML and NN 103 xii

15 xiii LIST OF FIGURES FIGURE TITLE PAGE Figure 1.0 Focus Area of Study 6 Figure 2.0 System Development Cycle of Waterfall Model 17 Figure 3.0 Process capture satellite image 24 Figure 3.1 Envi Interface 29 Figure 4.0 LST process 35 Classification process using maximum likelihood Figure 4.1 method 40 Overview of how remotely sensed data are turned Figure 5.0 into useful information 47 Figure 5.1 Spatial Subset (Subsetting by Images) 52 Figure 5.2 (a) Image before subset, (b) Image after subset 52 Histogram of the difference image value: (a)image Figure 5.3 before subset, (b)image after subset 55 Ocean masking: (a)masking into 0 and 1 data value, Figure 5.4 (b) Masking into 0 data value only 56 Cloud masking: (a) Image before mask, (b) Image Figure 5.5 after mask 58 Figure 5.6 Classification using Maximum Likelihood method 60 Figure 5.7 Temperature value of land cover histogram 62 Temperature value range for different land cover Figure 5.8 type 64 Comparison Landcover Classification Using (a) Figure 6.0 ROI and (b) Maximum Likelihood 76 Regions of interest (ROIs) Using by Maximum Figure 6.1 likelihood and Neural Network 77 Comparison Landcover Classification Using (a) Figure 6.2 ROI and (b) Neural Network 82 Comparison Landcover Classification Using (a) Figure 6.3 Maximum Likelihood and (b) Neural Network 86 Comparison Landcover Classification Using (a) Figure 6.4 Maximum Likelihood and (b) Neural Network 90 Figure 7.0 Main Page of Land Surface Temperature 96

16 Determination Application Second Page of Land Surface Temperature Figure 7.1 Determination Application 96 Third Page of Land Surface Temperature Figure 7.2 Determination Application 97 Fourth Page of Land Surface Temperature Figure 7.3 Determination Application 98 xiv

17 xv LIST OF ABBREVIATIONS LST TM OS UHI IFOV ML ENVI MATLAB ROI ARSM MACRES MOSTI - Land-surface temperature - Thematic Mapper - Operating Windows - Urban Heat Island - Instantaneous Field Of View - Maximum Likelihood - ENvironment for Visualizing Images - Matrix Laboratory - Regions of Interest - Malaysian Remote Sensing Agency - Malaysian Centre for Remote Sensing - Ministry of Science, Technology and Innovation TDRS GCP IR - Tracking Data and Relay Satellites - Ground Control Point - Infra-Red PSM2 - Projek Sarjana Muda 2 NN - Neural Network

18 1 CHAPTER 1 INTRODUCTION 1.1 Project Background Image processing is a technology that provides numerous benefits for the betterment of life. Remote sensing is one of technology used in image processing. The use of digital image processing for land temperature survey and mapping was initiated with the establishment of the National Remote Sensing Agency and Regional Remote Sensing Service Centres. Land-surface temperature (LST) is the thermal emission from the earth surface, including the top of the canopy for vegetated surfaces as well as other surfaces (such as residential landscape). LST plays an important role in the field of atmospheric sciences as it describes surface-atmosphere interaction between the atmosphere and the ground. Land surface temperature is an important parameter in determining the heat and moisture flow between the surface and the atmosphere. LST is also important in hydrologic, vegetation and decomposition processes. LST helps to control water exchange and surface heat with atmosphere.

19 2 This study used data recorded from a remote sensing satellite. Landsat platform use TM (Thematic Mapper) sensor. It is a multispectral scanning Earth surface. Images captured from such sensor usually has sharper spectral separation and greater image resolution. Other than that, it also has greater geometric accuracy. The TM data are scanned simultaneously by the TM sensor that is equipped with different spectral bands. Before the advent of this technology, land mapping information is done by using a plane that would cost. By using this technology mapping method can be done in a better way and save costs. Nature of technology requires data from satellite images to identify and use the land surface through the spectral response. 1.2 Problem Statements The study of LST mapping using remote sensing technology offers a new alternative over the convensional methods practiced elsewhere. This technology has not been widely used in Malaysia as compared to western countries. One of the main problems encountered in the conventional methods is that the LST data may contain variety of disturbance such as a noise. The disturbance of images will affect the accuracy of the temperature of the images. The traditional way to mapping process are use a plane and Sling Psychrometer at certain point on the ground. The ambient temperature variation only cover within a limited space. It needs a longer time to get the final result because the data need to be combined manually before getting the result. Involvement of plane required a big amount of cost. The

20 3 technique used does not give a good result because they could not prevent the images from environmental interference such as noise. Moreover, ground truth information of earth surfaces is hard to obtain due to the factors such as time, logistic and cost. The aim of the study is to derive derive LST information using remote sensing technology. 1.3 Objectives In order to achieve aim of the study, the specific objectives are: 1) To design a method to derive LST information. 2) To develop LST information. 3) To assess the accuracy of the LST map. 1.4 Scopes The Scopes of the study is divided into three part: software scope, area of study scope and user scope.

21 Software Scope All of the software in Table 1.0 is required in LST mapping process from start to end of the process of mapping. Software Operating Windows (0S) Software Processed Software Support Requirement Microsoft Windows 7 Ultimate Envi 4.5(Used to process and analyze geospatial imagery.) Matlab 2009a Database Support Microsoft Excel 2010 ( Use to represent the data.) Table 1.0: List of software required

22 Area of Study Scope Klang was chosen as the focus area of the study. Klang, Selangor is located at latitude: and Longitude: , 573KM2). Klang is the royal town and former capital of Selangor. The Klang River was separates Klang town into north and south sections. Klang is also the main port of Malaysia. On the east side, Klang borders with Shah Alam, on the north side borders with Kuala Selangor, on the west side borders with the Straits of Melaka and borders with Banting (Kuala Langat) to the south. Klang is booming and becoming the focus city rapid development and high growth of urbanisation and industrialisation. If unplanned development is practiced in Klang, various problems tend to exist. These urban transformations have altered the physical and natural environment. The urban surfaces absorb heat and will increase the temperature. This leads to Urban Heat Island (UHI) phenomenon. Therefore, a systematic urban planning is needed in order to avoid this phenomenon to occur. By using an appropriate Land Surface Temperature Determination the urban areas can be planned and developed systematically. Figure 1.0 show the location of Klang as the focus area of study. Spectral used in this area is band 1 until band 7. Landsat 5 TM was use as a platform in data processed and recorded. Landsat 5 is the fifth satellite of the Landsat program. It has a maximum transmission bandwidth of 85 Mbit/s. It was used at an altitude of mi (705.3 km). It takes some 16 days to scan the entire Earth. The Thematic Mapper (TM) is an advanced satellite compare to others. With the multispectral scanning capabilities, Earth resources sensor are capable to achieve higher image

23 6 resolution, sharper spectral separation, improved geometric fidelity and greater radiometric accuracy and resolution. TM data are processed in seven spectral bands simultaneously (Table 1.1). Band 6 senses thermal (heat) infrared radiation. Landsat can get night scenes in band 6. A TM scene has an Instantaneous Field Of View (IFOV) that capable to view image of 30m x 30m in bands 1-5 and 7 while band 6 has an IFOV of 120m x 120m on the ground. Landsat 5 (TM sensor) Wavelength(micrometers) Resolution (meters) Band Band Band Band Band Band Band Table 1.1 : Spectral range of bands and spatial resolution for the Landsat 5 TM sensor

24 Figure 1.0 : Focus Area of Study 7