(Luận văn) application of satellite image and geographic information system (gis) to build up the land cover change map for the management of taipei city

THAI NGUYEN UNIVERSITY UNIVERSITY OF AGRICULTURAL AND FORESTRY NGUYEN HUONG GIANG APPLICATION OF SATELLITE IMAGE AND GEOGRAPHIC INFORMATION SYSTEM TO BUILD UP THE LAND COVER CHANGE MAP FOR MANAGEMENT OF TAIPEI CITY BACHELOR THESIS an Lu va n Study Mode : Full-time : Environmental Science And Management ac th Major si : International Training and Development Center Batch : 2010 - 2015 ad o nl w Faculty l ul nf va an lu n oi m Thai Nguyen 30/09/2015 t i z z Thai Nguyen University of Agriculture and Forestry Degree program Bachelor of Environmental Science and Management Full name NGUYEN HUONG GIANG Student ID DTN1153180029 Application of satellite image and geographic Thesis Title information system (GIS) to build-up the land cover change for the management of Taipei city Ph.D, Associate Professor Tang-Huang Lin (National Central University, Taiwan) Associate Professor Chung-Pai Chang (National Central Supervisor(s) University, Taiwan) Associate Prof Dr Do Anh Tai ( Thai Nguyen University) Abstract: an Lu The main objective of this research is aimed at producing the land use and land va n cover change (LULCC) maps of the Taipei city using satellite image and th ac geographic information system (GIS) with a focus on urban development during si o nl w the period 2004 – 2014 The producer’s accuracy for both two images ranged from 89.4 % to 98.4 %, whilst the user’s accuracy ranged from 85.0 % to 99.4 % The of 2004 and 2014 were 94.6% and 90.4%, va an lu coefficient ad overall accuracies of 0.90 and 0.85, respectively with kappa A post-classification comparison l ul nf change detection algorithm was used to determine changes for land cover from n oi m t ii z z 2004 to 2014.The LULC from Landsat images map showed about 2883 (10.61%) area has been changed from vegetation to building The amount of building area increased by 4995 (18.39 % of the total area), while vegetation area decreased by 5075 (18.69 %) In addition, the brightness temperature map (BTM) from thermal infrared image observed from satellite can be utilized for assessing the impact of urbanization on thermal environment during the study period The result shows that the land surface brightness temperature (LSBT) increased 3.690 C from 2004 to 2014, indicating the more building area the higher temperature usually is By applying the satellite visible images associated with ArcGIS and ENVI software, the results of LULC and LSBT change detection can provide the reference for land management and environmental protection more efficiently Taipei city is the biggest industrial and commercial center in Taiwan The population reached 2,618,772 people in 2010 already Compared with 2009, the total population had been increased by 11,344 people (Taipei lookbook, 2010) The population growth and socio-economic an Lu development results in rapid increasing transportation and urban expansion to va n suburban regions The effects of urbanization on local weather and climate change th ac resulted in a remarkable increase in mean temperatures from 2004 to 2014 si o nl w According to the assessments, the strategy can be proposed for the management and protection environment in Taipei city, Taiwan ad l ul nf va an lu n oi m t iii z z Taipei city, Land cover change map, ArcGIS software, Keywords ENVI software, urbanization, Land surface brightness temperature Number of pages 64 Date of submission September, 30, 2015 an Lu n va ac th si ad o nl w l ul nf va an lu n oi m t iv z z ACKNOWLEDGEMENT First and foremost, I wish to express our sincere thanks to Center for Space and Remote Sensing Research (CSRSR) of National Central University (NCU), Taiwan for providing me all necessary facilities, skills and knowledge to complete this bachelor thesis Furthermore, I want to deeply thank our principal research adviser Assoc Prof Tang-Huang Lin, Prof Chang Chung-Pai and Assoc Prof Do Anh Tai guided me wholeheartedly when I study for my project Especially thankful to International Training Center – Thai Nguyen University of Agriculture and Forestry has facilitated for me the chance to come here to study and get more knowledge exchange Finally yet importantly, I take this opportunity to record my sense of gratitude to my families and friends who encourage and backing me unceasingly an Lu n va ac th Thai Nguyen, 30/09/ 2015 si ad o nl w Author l ul nf va an lu Nguyen Huong Giang n oi m t v z z TABLE OF CONTENTS LIST OF FIGURES LIST OF TABLES …2 LIST OF ABBREVIATIONS PART I INTRODUCTION 1.1 Research rationale 1.2 Research objectives 1.3.The requirement 1.4 The significance: PART II LITERATURE REVIEW 2.1 Theoretical basis 2.1.1 The land cover 2.1.2 The land covers change 11 2.1.3 Geographic information system (GIS) 12 2.1.4 Arcgis software 14 an Lu 2.1.5 Remote Sensing(RS) 15 n va 2.1.6 The Landsat program 17 ac th 2.1.7 ENVI software 19 si 2.1.8 Brightness Temperature 20 w 2.2 Practical basis 21 o nl 2.2.1 The research in the world 21 ad va an lu 2.2.2 The research in Viet Nam 24 PART III.MATERIALS AND METHODS 28 l ul nf 3.1 Materials 28 n oi m 3.1.1 The objects and scope of research 28 t vi z z 3.1.2 The content of research 28 3.2 Methods build-up land cover change map in period 2004-2014 28 3.2.1 Data Collection 30 3.2.2 Image Preprocessing 31 3.2.3 Image classification 35 3.2.4 Post Classification 37 3.3 Methods Build-up Brightness Temperature Map 40 3.3.1 Data collection of Landsat image thermal band 41 3.3.2 Conversion of the Digital Number (DN) to Spectral Radiance (Lλ) 42 3.3.3 Conversion to At-Satellite Brightness Temperature 43 3.3.4 Transfer the temperature value from Kelvin unit to Celsius unit 43 PART IV RESULTS 44 4.1 The natural conditions and socioeconomic in research area 44 4.1.1.Natural conditions 44 4.1.2 Socioeconomic conditions in Taipei city 46 4.2 Classification accuracy assessment of Taipei city in 2004 and 2014 48 4.3 Land Cover Maps 51 an Lu 4.3.1 Landsat classification area statistics for 2004 and 2014 51 4.3.2 Land cover Map for Landsat TM (2004) 52 va n 4.3.3 Land cover Map for Landsat TM (2014) 53 th ac 4.4 Land cover change map from 2004 to 2014 55 si w 4.5 Brightness temperature map (BTM) for Taipei city 2004 and 2014 56 o nl PART V DISCUSSION AND CONCLUSION 59 ad 5.1 Discussion 59 va an lu 5.2 Conclusion 60 l ul nf REFERENCES 62 n oi m t vii z z LIST OF FIGURES Figure 2.1: Remote Sensing procedures .17 Figure 3.1(a): Satellite images cover the research area in 2004 31 Figure 3.1(b): Satellite images cover the research area in 2014 31 Figure 3.2(a): Composite bands 4, 3, for LS-5 TM image; July 12th 2004 32 Figure 3.2(b): Composite bands 5, 4, for LS-8 OLI/TIRS Pre-WRS2 image; August 25th, 2014 33 Figure 3.3: Landsat images shown in True and False composite colors 35 Figure 3.4: The classification image after we used the tool filtering 38 Figure 3.5: The Landsat thermal band image of Landsat and Landsat 42 Figure 3.6: The formula calculate spectral radian of Land sat 5(2004) and Landsat (2014) 42 Figure 3.7 : The formular calculate the At-Satellite Brightness Temperature 20042014 43 an Lu Figure 4.1: The Geographical location map of Taipei city 45 n va Figure 4.2: The classified images of land cover map for 2004 52 ac th Figure 4.3: The statistics percentage of land cover type classification in 2004 52 si Figure 4.4: The classified images of land cover map for 2014 53 w Figure 4.5: The statistics percentage of land cover type classification in 2014 54 o nl ad Figure 4.6: Land cover change of Taipei city in 2004-2014 55 va an lu Figure 4.7 : Brightness temperature map of Taipei city in 2004 57 l ul nf Figure 4.8 : Surface brightness temperature (SBT) map of Taipei city in 2014 57 n oi m t z z LIST OF TABLES Table 2.1: Level of the classification 10 Table 2.2 : Parameters of ETM Landsat (Landsat 5) 18 Table 2.3: Parameters of LDCM Landsat (Landsat 8) 19 Table 3.1: The information of Landsat image 30 Table 3.2: Land cover classification scheme 39 Table 3.3: K1 and K2 Values in Landsat 8and Landsat5 images 41 Table 3.4: Rescaling Factor in Landsat image and Landsat images 41 Table 4.1: Accuracy assessment of LULC classification in 2004 and 2004 48 Table 4.2 : Accuracy assessment of classified land cover change in 2004 49 Table 4.3: Accuracy assessment of classified land cover change in 2014 50 Table 4.4 : Summary of classification area with Landsat data in 2004 and 2014 51 Table 4.5 : Statistical fluctuations of land cover change in the period 2004 - 2014 55 an Lu n va ac th si ad o nl w l ul nf va an lu n oi m t z z LIST OF ABBREVIATIONS BTM Brightness temperature map CASI Compact Airborne Spectrographic Imager DN Digital Number ENVI The Environment for Visualizing Images EROS Earth Resources Observation and Science ESRI Environmental Systems Research Institute FLIR Forward Looking InfraRed GDB Create Geodatabases GIS Geographic Information System IDL Interactive Data Language Indian Remote Sensing LCC Land-Cover Changes LCCS Land Cover Classification System Lidar Light Detection and Ranging LULCC Land-use and land-cover change an Lu IRS n va Multispectral Electro-optical Imaging Scanner ac th MEIS-II Marine Observation Satellite NIR Near Infrared NDVI The normalized vegetation index NASA National Aeronautics and Space Administration OLI The Operational Land Imager ROI Region of Interest si MOS ad o nl w l ul nf va an lu n oi m t z z classification were turned out to be better than expected The percentage of categories classified have accuracy relatively high such as: The percentage of water bodies , vegetation and building occupied 96.08% , 98.4% and 87.75% respectively Table 4.3 : Accuracy assessment of classified land cover change in 2014 Reference ( Percentage) Classified (Percentage) Water Bodies Vegetation Building Total Water Bodies 89.42 0.15 1.48 8.17 Vegetation 3.79 98.05 9.04 17.81 Building 6.79 1.80 89.48 74.02 Total 100 100 100 100 Overall Accuracy = 90.3596 Kappa Coefficiention = 0.8545 By flowing the table 4.3, the overall accuracy of classification image dated 2014 was 90.36% and the Kappa coefficient was 0.85% The accuracy matrix of categories an Lu classified have accuracy relatively high, such as The percentage of water bodies, va n vegetation and building occupied 96.08%, 98.4% and 87.75%, respectively ac th si ad o nl w l ul nf va an lu n oi m t 50 z z 4.3 Land Cover Maps 4.3.1 Landsat classification area statistics for 2004 and 2014 Table 4.4: Summary of classification area with Landsat data in 2004 and 2014 Years Class 2004 2014 Relative change Area(ha) Area(%) Area(ha) Area(%) Area(ha) Area(%) Water Bodies 726.48 2.67 805.95 2.97 79.47 0.3 Building 9333.36 34.35 14328.99 52.74 4995.63 18.39 Vegetation 17110.16 62.98 12035.06 44.29 -5075.1 -18.69 Total 27170 100 27170 100 The land cover classification results are summarized for the years 2004 and 2014 in Table 4.4 From 2004 to 2014, water bodies areas increased 79.47 (0.3%) with quite low percentage Besides that, building areas raised relatively with number 4995.63 (18.39%) On the other hand, vegetation areas decreased 5075.1ha an Lu (18.69%) Therefore, we can realized that having a change area between two classification is building and vegetation As seen in Table 4.4, the area well qualified va n for building area was 9333.36 in 2004, while the building area increased about th ac 18.39% and reached up to 14328.99 in 2014 On the other hand there was a 5075.1 si w decreased in vegetation areas (decreased from 17110.16 to 12035.06 ha) during ad o nl the period between 2004 and 2014 The increase in building areas (4995.63 ha) and the decrease in forest areas (5075.1 ha) are approximately the same These given data va an lu expressly state that the increase in building areas mostly result in processing built-up n oi m in the region l ul nf urban which means some vegetation areas were removed and converted to the building t 51 z z 4.3.2 Land cover Map for Landsat TM (2004) Figure 4.2: The classified images of land cover map for 2004 an Lu n va ac th si ad o nl w l ul nf va an lu Figure 4.3 : The statistics percentage of land cover type classification in 2004 n oi m t 52 z z As indicated in Table 3.2, the classification scheme used in this study comprised of three land cover categories, namely building, vegetation and water bodies Figure 4.2 and figure 4.3 shows the land cover for the classified Landsat (2004), and its corresponding statistics percentage of land cover type classification As indicated by the statistics and as it can be seen in the image, nearly two thirds (2/3) of the study area was covered by forests, i.e 62.98 % representing 17110.16 During this period, thick vegetation surrounded building areas and concentrated mainly on the Eastern, with some significant vegetation cover within settlement area.The building area during this time occupied only about 23.89 % (180.73 Km2), and were concentrated along the Western portion of the image nearn river The river interspersed the settlement areas, were the major water bodies alongside scattered water ponds, accounting for 2.67% (726.48 ha) of the study area 4.3.3 Land cover Map for Landsat TM (2014) an Lu n va ac th si ad o nl w l ul nf va an lu n oi m Figure 4.4: The classified images of land cover map for 2014 t 53 z z Figure 4.5: The statistics percentage of land cover type classification in 2014 Figure 4.4 shows the land cover classification map from the Landsat image of 2014 A close look at the image reveals that there was a significant intrusion of building areas in to the surrounding vegetation and the reduction of the scattered vegetation cover in settlement areas that was present in the 2004 classification image an Lu (Figure 4.2) This left the vegetation cover to have made up 44.29 % (12035.06 ha) of n va the total area, as shown by the histogram in figure 4.5 With continuous growth ac th registered since 2004, the building area was more significant than ever before in 2014, covering up to 52.74 % ( 14328.99 ha) The fraction of water bodies made up about si ad o nl w 2.97% (805.95 ha) of total area over Taipei city l ul nf va an lu n oi m t 54 z z 4.4 Land cover change map from 2004 to 2014 Figure 4.6: Land cover change of Taipei city in 2004-2014 Table 4.5 : Statistical fluctuations of land cover change in the period 2004 - 2014 Areas( hectares) 663.66 % Total areas 2.44 17.4 0.064 47.97 0.18 Vegetation to Water 110.25 0.4 Vegetation 14109.78 51.93 2883.05 10.61 30.51 0.11 an Lu Changing of Land Cover Type Water va Water to Vegetation n ac th Water to Building si Building va an lu Building to Vegetation ad Building to Water o nl w Vegetation to Building 201.81 0.74 9104.04 33.51 l ul nf n oi m t 55 z z According to the table 4.5 about statistical fluctuations of land cover change in period 2004-2014 Thefore, we can analysis of land cover change based on "from-to" categories between the 2004 and 2014 images revealed that changing from vegetation to building occupied the highest was 2883.05 (10.61%) , followed by changing from water to building occupied 47.97 (0.18%) and the fewest were changing from water to vegetation occupied 17.4 (0.064%) From results analysis the figure 4.6, the land cover change classified images were visually inspected for changes from 2004 to 2014 The figure clearly showed the land cover classes from vegetation to building that have changed in this study period of about a decades with increase of the building area and decrease of the vegetation classes been very prominent This is highly related to the issue of urbanization speed The post-classification comparison gives a better idea of the changes that occurred within a period of 10 years by visually looking at the images and inspecting the areas where settlements has expanded and deforestation has occurred an Lu 4.5 Brightness temperature map (BTM) for Taipei city 2004 and 2014 Through the two image about BTM from 2004 to 2014, it can be seen that the n va ac th major temperature range from 0C - 35 0C, and it divided six groups and si corresponding with each group is other colours For example, in the group highest w temperature was 300 C-350C show by red color, the group lowest temperature was 00C ad o nl -150C show by green color va an lu The overall, the high surface brightness temperature (SBT) of more than 27 0C l ul nf was focused in the metropolitan of the study area, where have building such as n oi m industrial, residential and transportation areas were mostly found The areas with t 56 z z surface brightness temperature lower than 250C were traced in the vegetation regions such as river area and cultivable land in the plains The lowest category of less than 150C was seen in the forest highly elevated regions with dense vegetation Figure 4.7 : Surface brightness temperature (SBT) map of Taipei city in 2004 an Lu n va ac th si ad o nl w va an lu l ul nf Figure 4.8 : Surface brightness temperature (SBT) map of Taipei city in 2014 n oi m t 57 z z The results show that the mean SBT in 2004 and 2014 are 22.170 C and 25.860 C, respectively By comparing between 2004 and 2014 then we can see the increasing SBT clearly in 2004 the mean temperature rises about 3.690C since 2004, which may cause form the increase of building area With the status present, we need to propose the effective solution in order to suppress the increasing SBT during the period of urbanization an Lu n va ac th si ad o nl w l ul nf va an lu n oi m t 58 z z PART V DISCUSSION AND CONCLUSION 5.1 Discussion In order to prevent the impacts of urbanization on land cover changes and surface temperature increase, some methods and policies of environmental management should be proposed Therefore we need to discuss the effective solution such as: + Local government and private sectors should invest in educational and research activities to constantly assess, monitor, and create an up-to-date database on the status of land and natural resources There is need to raise public awareness of the importance and limitedness of land resources This will help facilitate in land related decision-making and improve the management of land resources + Proper and realistic policies and strategies that focused on environmentally benign activities should be put in place and implemented for the sustainable management of land and its resources These policies and strategies need not only an Lu conserve the land resources but protect the rights of populace n va + The human caused stress on the vegetation is reaching its worst point; people th ac and their related socio-economic activities are the major source and cause for land si cover change on the study area As a result of these activities, the wild life habitat and w ad o nl biodiversity of the area is affected greatly Parks and nature reserves should be free from permanent human settlements through resettlement programs and other l ul nf va an lu destructive human interferences n oi m t 59 z z 5.2 Conclusion Based on the analysis and discussion of results in this study, the following conclusions can be summarized This research demonstrates the ability and applications of Remote Sensing and GIS in land-cover change detection Attempt was made to classify as accurate as possible for land cover classes as they change through time with emphasis Three classes of LULC were built-up class as they are a combination of anthropogenic activities which affect other categories Overall, the method of change detection yields acceptable results and the potential for using Remote Sensing Technique in detecting land-cover changes in other parts of the country The results of this study demonstrated that Landsat classification images can be used to produce accurate and reliable change maps, and also can be successfully used to predict the likely trend of the future land cover changes In this research, the overall accuracies of 2004 and 2014 about 94% and 90% with kappa coefficient of 0.90 and 0.85, respectively an Lu The usage of satellite imagery is much appreciated, since it has the advantage va n of not only showing the statistical information for the changes but also visualizing the th ac areas that have or are expected to undergo changes In this study, the conversion of si o nl w vegetation areas to building is the most significant transformation both in terms of magnitude and with regards to the threat and dangers that it poses to the environment ad va an lu From the results, it can be seen that clear patterns highlighting urbanization activity in the metropolitan area have emerged and increased speed rapidly The results further l ul nf indicated that the study area (Taipei city) has changed, and the changes mostly n oi m t 60 z z affected the urban building and the vegetation areas in terms of the magnitude of land cover area Overall, urban building increased of 18.39 % (4995.63 ha), and most increase of the pixels of this was created from the vegetation land cover categories during the period of the study Therefore, the vegetation land cover categories decrease of 18.69% (5075.1ha) an Lu n va ac th si ad o nl w l ul nf va an lu n oi m t 61 z z REFERENCES Bikash, B., & Anthony, V.(2015) Tracking Land Use/Land Cover Dynamics in Cloud Prone Areas Using Moderate Resolution Satellite Data Chander, G., & Markham, B.(2003) Revised Landsat-5 TM radiometric calibration procedures and post calibration dynamic ranges Geoscience and Remote Sensing, IEEE Transactions on, 41,p 2674 – 2677 Duong, N.D., Thoa, K., & Hoan, N.T (2005) Monitoring of forest cover change in Tanh Linh district, Binh Thuan province, Vietnam by multi-temporal Landsat TM data Duong, N D (2006) Study land cover change in Vietnam in period 2001-2003 using MODIS 32 days composite Ehlers, M., Welch, R., & Ling, Y (2004) GIS and Context Based Image Enhancement Proceedings of the XXth International Congress of ISPRS, Istanbul, Turkey, IAPRS XXXV/B4, pp 397-402 Ellis, E (2013) Land-Use and Land-Cover Change an Lu http://www.eoearth.org/view/article/154143 Retrieved: 16 – 05 – 2015 n va Geoffery, J M (1996) Geographical Information Systems Applications to Marine ac th Fisheries, pp 127-146 si Hien, N.T.T (2013) Assessing Land Use and Land Cover Change: A Case of Tien w Yen District, Quang Ninh Province from 2000 to 2010 o nl ad James, R A., Ernest, E.H., John, T R., & Richard, E.W (1976) A Land Use And va an lu Land Cover Classification System For Use With Remote Sensor Data United l ul nf States Government Printing Office, Washington n oi m t 62 z z Kristina, S., Francisco, O., & David, R.M (2002) Introduction to ArcGIS Center for Research in Water Resources ,University of Texas at Austin Liu, H (2010) The Basics of ENVI and the Nature of Digital Images Texas A&M University Lakshmi, P & Sreenivasulu (2014) International Journal of Scientific and Research Publications, Volume 4, Issue 6, June 2014 ISSN 2250-3153 National Statistics, Republic of China – Latest Indicators Directorate General of Budget, Accounting and Statistics Oštir, K ( 2006) Daljinsko zaznavanje Založba ZRC, p 250, Ljubljana Pouncey, R., Swanson, K., and Hart, K.(1999) ERDAS field Guide Fifth Edition, Revised and Expanded, ERDAS Field Guide Phi, N.N (2009) Applications remote sensing monitoring land urban change at Vinh City, Nghe An Province Pal Nikolli (2010) Tirana University, Faculty of History and Philology, Department of Geography, Tirana, Albania Land cover classification in Albania, pp 414 Sokal, R (1974) Classification: purposes, principles, progress, prospects Science an Lu 185(4157), pp 1115-1123 n va Santhosh, B (2011) Geometric Correction in Recent High Resolution Satellite ac th Imagery: A Case Study in Coimbatore, Tamil Nadu Postgraduate and si Research of Computer Science, Dwaraka Doss Goverdhan Doss Vaishnav o nl w College, Chenna ad Simon, F F., & Gregory, N T.(2013) Using satellite data to monitor land-use land- va an lu cover change in North-eastern Latvia l ul nf n oi m t 63 z z Southeast Asia Travel Climate Tour (2015) https://www.worldspree.com/escortedsoutheast-asia-tours/amazing-vietnam-tour-2015/southeast-asia-travelclimate.htm Retrieved 2015-07-13 Taipei City Has Second-highest Per Capita GDP in Asia.(2009) China Economic News Service Taiwan's tourism revenue on the rise.(2015) Focus Taiwan News Channel Vinh, T.Q.(2009) Using Satellite Data for Mapping Land Cover Factor (C) in Soil Erosion Research in Tam Nong District Phu Tho Province an Lu n va ac th si ad o nl w l ul nf va an lu n oi m t 64 z z