the land cover mapping of dong hy district, thai nguyen province using satellite images and gis

THAI NGUYEN UNIVERSITY OF AGRICULTURE AND FORESTRY PROGRAM ADVANCED EDUCATION  Student name: Nguyen Binh Minh Student ID: 1053060032 K42 - AEP THE LAND COVER MAPPING OF DONG HY DISTRICT, THAI NGUYEN PROVINCE USING SATELLITE IMAGES AND GIS Supervisor: Msc Nguyen Van Hieu Thai Nguyen 15th January, 2015 ABSTRACT Land use/cover change mapping is one of the basic tasks for environmental monitoring and management This research was conducted to analyze the land use and land cover changes in Dong Hy district, Thai Nguyen province In recent years, a variety of change detection techniques have been developed The data sources used in this study were Landsat and Landsat images taken in November 2004, and December 2013, respectively By using ArcGIS and ENVI software and remote sensing data, a supervised classification was performed based on fusion data from a composite image of the bands Using this output, available secondary data together with field data in order to perform a Maximum Likelihood supervised classification Six classes were classified, namely forest, water, mineral, residential, traffic, rice – crops and water With overall accuracy 96.4092% and kappa = 0.9502 in 2004 , overall accuracy 96.2690% and kappa coefficient = 0.9529 in 2013 After conducted, we have: - Land cover map of Dong Hy district in 2004 and 2013 - Land cover changes map of Dong Hy district in period 2004 – 2013 With the results achieved, we can realize the remote sensing and GIS technology is effective method for high accuracy, cost savings in the classification and analysis of land cover changes ACKNOWLEDGEMENT Approved by International Training Center – Thai Nguyen University of Agriculture and Forestry, with the enthusiastic help of Dong Hy district, the teachers, and peoples at Center for Foreign Language and Applied Informatics, I have done to implement the research: “The land cover mapping of Dong Hy district, Thai Nguyen province using satellite images and GIS” Through the implementation process I have gained much useful knowledge as well as certain results I would like to express my special thanks of gratitude to my teacher Msc Nguyen Van Hieu who gave me the golden opportunity to this wonderful project, which also helped me in doing a lot of Research and i came to know about so many new things I am really thankful to them Secondly i would also like to thank my parents and friends who helped me a lot in finishing this project within the limited time I would also like to thanks the officers and staffs of the Dong Hy district who enthusiastically communicated word experience and helped me a lot in the supply of data for my research to create conditions for I can complete this research In addition, I would like to thank family, friends and relatives who were always at my side to encourage and help me in the learning process as well as during the time I performed research Again, I sincerely thank! Student Nguyen Binh Minh Thai Nguyen, January ,2015 LIST OF ACRONYMS GIS Geographic information systems NDVI Normalized Difference Vegetation Index RS Remote sensing SPOT System Pour l'Observation de la Terre IRS Indian Remote Sensing Lidar Light Detection and Ranging MOS Marine Observation Satellite SeaWiFS Sea-viewing Wide-Field-of View Sensor FLIR Forward Looking InfraRed RADAR RAdio Detection And Ranging USLE Universal Soil Loss Equation GDP Gross Domestic Product VND Vietnamese dong ROI Region Of Interest LISTS OF TABLES Table Name Table Land cover system Table Landsat satellite system Table Parameters of ETM Landsat ( Landsat 5) Table Parameters of LDCM Landsat (Landsat ): Page 13 14 15 Table Statistics sample collection 22 Table 6: Economic development in Dong Hy district period 2011 – 23 2013 Table The norms of income 30 Table The information of Landsat 33 Table 9: Recognizing the features on images and fields 34, 35 Table 10 Results of the accuracy evaluation in 2004 40 Table 11: Results of the accuracy evaluation in 2013 41 Table 12 Statistical fluctuations of land cover in the period 2004 - 45 2013 LIST OF FIGURES I INTRODUCTION 1.1 Background 1.1 Purposes 1.2 Requirements 1.3 Signification II LITERATURE REVIEW 2.1 Theoretical basis 2.1.1 Definitions of land cover 2.1.1.1 Land cover 2.1.1.2 Normalized Difference Vegetation Index 2.1.2 Geographic information system (GIS) 2.1.2.1 Geographic information system (GIS) 2.1.2.2 ArcGIS software 2.1.3 Remote sensing (RS) 2.1.3.1 Remote sensing (RS) 2.1.3.2 The Land sat program 12 2.2 Practical basis 16 2.2.1 The research in the world 16 2.2.2 The research in Viet Nam 18 3.1 Object, scope and time of research 21 3.1.1 The objects of research 21 3.1.2 The scope 21 3.2 Content of research 21 3.3 Methodology 21 3.3.1 Data collection 21 3.3.2 Field trips method 22 3.3.3 Building the land cover changes map 23 3.3.4 Normalized difference vegetation index (NDVI) 24 3.3.5 Accuracy assessment and image processing after classified 24 3.3.6 Building map 25 IV RESULT AND DISCUSSION 26 4.1 Evaluating the natural conditions and socioeconomic in research area 26 4.1.1 Natural conditions 26 4.1.1.1 The geographic location 26 4.1.1.2 The topography and geomorphology 28 4.1.1.3 The climate and hydrology 28 4.1.1.4 Natural resources 28 4.1.2 Socioeconomic conditions 29 4.1.2.1 Economic development status 29 4.1.2.2 The Population, labor and employment 30 4.1.2.3 The culture and society 31 4.1.2.4 The infrastructure status 31 4.2 The process of current status land cover mapping 32 4.2.1 Data preparation 32 4.2.1.1 Data collection 32 4.2.1.2 Data description 33 4.3 Analyze remote sensing image, determine land cover in Dong Hy district 33 4.3.1 Image interpretation 33 4.3.2 The process of calculate NDVI 36 4.3.2.3 Remote sensing image classification 38 4.3.2.4 Evaluating the accuracy after classification 40 4.3 ArcGIS application, editting current status land cover 42 4.3.1 Building current status land cover map 42 4.3.2 Building fluctuations map 43 4.3.3.3 Analysis of fluctuation 45 V CONCLUSION AND RECOMMENDATION 47 5.1 Conclusion 47 5.2 Recommendation 48 VI REFERENCES 49 I 1.1 INTRODUCTION Background Land cover is all the material composition of natural and artificial cover on the earth’s surface includes: vegetation, the constructions of human, soil, water, sandy soil… Land cover show the current status Over time, land cover is continuous change under strong impact of disasters, human – That is the economic – Social development activities Research mapping land cover using remote sensing and GIS technology helps to shorten the time compared to the built maps technologies previously and it is important contributions in the management of natural resources, assess the current state of vegetation With these pressures, land and land cover are constantly fluctuating with the development of society This is a special resource can exploitation and use but can not increase in quantity Therefore, the monitoring, research, management and the use of natural resources is an effective and reasonable Remote sensing technology is increasingly widely used in many sectors, fields of meteorology - hydrology, geology, from environment to agriculture - forestry fisheries, including monitoring changes in the types of land cover with high accuracy, which can help managers have more resources to monitor land-use change This is considered as one of the solutions for the posed problems On the other hand, this method has not been tested application in the area of Dong Hy district So, research " The land cover mapping of Dong Hy district, Thai Nguyen province using satellite images and GIS " is performed 1.1 - Purposes Research overview of the land cover map, satellite images and geographic information systems (GIS) - Research on remote sensing and GIS technology in mapping land cover work - Research on the spectral properties of natural objects - Develop process mapping land cover by remote sensing and GIS technology - Assess the current state of the land cover in Dong Hy District, Thai Nguyen Province 1.2 Requirements - Adequate the data of natural condition, socioeconomic and spatial data - Classify and handling the data collected - The result of evaluating the current state of the land cover - Proficiency in using GIS software to mapping data and analyze the data 1.3 Signification Mapping land cover using remote sensing and GIS technology help to shorten the time compare to the built map technologies previously - The significance in learning and research: to learn the research methods, evaluating the changes in land cover - The significance in reality: Applying the knowledge on reality combine with collect and analyzing data Assessing the changes in land cover, providing information to community in the research area The map below illustrate the calculation result NDVI of in Dong Hy district in 23rd November 2004 As can be seen that the areas have high NDVI value mostly in the north, the center, and the east of Dong Hy district ( NDVI> 0.55) The areas with lower NDVI value is the west and south areas ( NDVI< 0.55) The highest NDVI value is 0.666667 and the lowest NDVI value is 0.254237 Figure 9:NDVI Map of Dong Hy in 2004 This is the NDVI map of Dong Hy district in 18 December 2013 The NDVI value is markedly changed From the map, we can see that the NDVI value is quite high The highest NDVI value is 0.589906 and the lowest NDVI value is -0.225976 Over years of change, we can see the vegetation cover has been enhanced Figure 10:NDVI Map of Dong Hy in 2013 37 4.3.2.3 Remote sensing image classification Firstly, data obtained included individual spectral channels we use ENVI software to open satellite image Then we combine spectral band we use Basic Tools/ Layer Stacking Figure 11: Layer stacking Secondly, because the research area is only one part of the picture so it should be conducted cropping A vector file containing the boundary area of Dong Hy district are used to cut the study area Figure 12: Area image 38 Next, we use ROI Tool to take sample After that, we find tool in menu Classification/Supervised/Maximum likelihood to classify and we have result: Figure 13: ROI tool After classify we use Classification/ Post Classification/ Majority/Minority Analysis to combine sporadic pixels Finally, click on Classification/ Post Classification/ Classification to Vector to output data 39 4.3.2.4 Evaluating the accuracy after classification When we already classify, we will evaluate the accuracy Click on menu on Classification/ Post Classification/Confusion Matrix/Using Ground Truth ROIs And we have results of the evaluation accuracy is expressed specifically in the tables By which, the result in 2004 with overall accuracy 96.4092% and kappa = 0.9502, the result in 2013 with overall accuracy 96.2690% and kappa coefficient = 0.9529 Table 10: Results of the accuracy evaluation in 2004 Unit: % Classes Forest Water Rice - Crops Minerals Residential Traffic Total Forest 100 0.91 0 0 18.02 Water 99.09 0 0 7.38 Rice - Crops 0 94.88 1.15 20 44.51 Minerals 0 0.29 100 0 14.30 Residential 0 0.88 97.13 14.29 12.20 Traffic 0 3.95 1.72 65.71 3.59 Total 100 100 100 100 100 100 100 40 Table 11: Results of the accuracy evaluation in 2013 Unit: % Classes Forest Water Rice - Crops Minerals Residential Traffic Total Forest 99.63 0 0 21.57 Water 98.88 0 0 14.14 Rice - Crops 0.37 1.12 95.15 0 6.33 28.83 Minerals 0 100 0 10.46 Residential 0 4.85 99.45 25.32 20.61 Traffic 0 0 0.45 68.35 4.39 Total 100 100 100 100 100 100 100 From the tables, we can make flowing comments: - Overall accuracy and kappa coefficient is high, can be trusted - The level of omission errors of traffic class is quite high ( about 34% in 2004 and 31% in 2013) The remaining layers are very low or equal zero Image data in 2004 have omission errors higher due to differences in period of time - Traffic class and Residential class are often misclassified due to the similarities in their spectral values with other classes 41 4.3 ArcGIS application, editting current status land cover 4.3.1 Building current status land cover map After classify, we use ArcGIS software to editing We right click on class already classify > Properties > Symbology > Categories > Unique values We use Value Field “Class_name” and then select right color Figure 14 (a) and (b): Edit and choose color After adding grid, legend, north arrow, title we have land cover map in 2004 and 2013 Figure 15 : Land cover map of Dong Hy district in 2004 and 2013 42 4.3.2 Building fluctuations map Right click on class already classify > Open Attribute Table We add field “Sign” and number Forest into group 1, Minerals into group 2, Residential into group 3, Rice – Crops into group 4, Traffic into group 5, Water into group To Join classes data in 2004 and 2013 we use Union tool Arctoolbox /Analysis/Overlay/Union Figure 16 : Attribute table Figure 17 : Union tool On layer have joined data, right click - select Open Attribute Table then adding a field with name “20042013” Right click on that field , select Field Calculator > add [Sign 2004] “to” [Sign2013] and OK Figure 18 : Field calculator 43 Finally, we have fluctuations results To display it clearly We right click on class already calculate > Properties > Symbology > Categories > Unique values We use Value Field “2004_2013” and then select right color Figure 19: Display classes F stands for Forest M stands for Minerals R stands for Residential RC stands for Rice – Crops T stands for Traffic W stands for Water Figure 20 : Land cover change map of Dong Hy district (2004 – 2013) 44 4.3.3.3 Analysis of fluctuation - From the land cover map and the statistical results of 2004 and 2013, conducted layer stacking to get land cover change in the period 2004 - 2013 Results of statistical fluctuations are shown as table Table 12: Statistical fluctuations of land cover in the period 2004 - 2013 (a) Unit: Classes Forest Water Rice - Crops Traffic Minerals Residential Forest 19762.92 11.78 390.78 2.91 9.09 32.16 Water 35.82 861.53 56.43 11.97 14.76 63.99 Rice - Crops 5059.46 52.74 12458.54 163.98 30.78 26.76 Traffic 846.87 6.93 696.09 870.84 18.81 226.10 Minerals 24.19 12.42 96.03 47.28 396.65 Residential 22.19 0.9 427.77 4.66 0.0990 (b) 1.98 631.49 Unit : % Classes Forest Water Rice – Crops Traffic Minerals Residential Forest 85.36 1.24 2.76 0.26 1.9 3.27 Water 0.15 91.04 0.4 1.07 3.13 6.51 Rice – Crops 13.21 5.57 88 14.75 6.54 2.72 Traffic 1.06 0.7 4.92 78.33 23.01 Minerals 0.1 1.31 1.19 4.25 84.35 0.2 Residential 0.12 0.14 3.03 1.34 0.08 64.29 45 Based on these results, we can see Forest, Rice – Crop, Water, Minerals is the most stable classes (more than 80% of the area unchanged) , while the Residential class, are change (have over 30% of change) For Forest class, about 85.36 % of the area is kept constant in the period 2004 - 2013, the rest was changed into other classes, such as Rice - crops (13.21%), Traffic (1.06 %), Residential (1.777%) For Rice - crops, approximately 88% of the area is kept constant in the period 2004 2013, the rest area lost was changed into Traffic, Forest, and Residential For Residential class, only 64.29% of the area is almost stable in the period 2004 2013, most of the area lost was changed into Traffic 46 V CONCLUSION AND RECOMMENDATION 5.1 Conclusion Dong Hy district, with the rate of change quite quickly Therefore, monitoring changes of land cover will give us correct information on the current status of land cover types, as well as its variations It is the change of scale, size and trends This is the scientific basis for making policies about land management effective and rational, be a premise for land use management Remote sensing technology combined with GIS give us high effectively and objectively assess the change in land cover The experimental results also indicate, the combination of remote sensing and GIS technology is very useful to determine the area of fluctuations and trends of each object Research has applied remote sensing technology combined with GIS for building land cover map of Dong Hy district in 2004, 2013 and land cover change map of Dong Hy period from 2004 to 2013, helping the planners, management of resources and environment, urban planning can assess more accurately the current status in the area Besides that, the satellite data often fails also caused many difficulties for the process of interpretation The use of satellite data in building land cover map is relatively simple and quite quickly, if this research is invested and widely used It will save the cost, effort and time, the results obtained equivalent or even ahead to the method of measurement and statistics in the field 47 5.2 Recommendation Due to the limited time and resources so the research only build land cover map, land cover changes map in 2004 and 2013 To reach result have high value and use as a source of data for policy planners, should use multiple images at the moment more and more Should combine multiple classification methods and other types of remote sensing data to make the interpretation for better results 48 VI REFERENCES Article in collective work: [1] G Sreenivasulu, 2013, Analysis on Land Use/Land Cover Using Remote Sensing and GIS – A Case Study In and Around Vempalli, Kadapa District, Andhra Pradesh, India, International Journal of Scientific and Research Publications [2] Ha Bui Nguyen Lam, 2011, Estimating Biomass of the canopy of leaves by using Satellite data ALOS AVNIR – 2, workshop nationwide GIS application 2011 [3] Hung Tran, Loi Pham Quang, 2008, Practical Guide: Processing and analysis of remote sensing data with software ENVI, GeoViet Company [4] Kham Duong Van, 2007, Use of multi data Remote sensing data to evaluate vegetation indices changes of the land cover and some analysis of crop and rice state in red river and Cuu Long river deta, Collection of scientific works - Scientific Meeting of Geography and Land Administration [5] Long Vu Huu, 2011, Using MODIS satellite images to research seasonal crops, building current status and land cover changes map in Hong river delta, workshop nationwide GIS application [6] M Harika et al., 2012, Land use/land cover changes detection and urban sprawl analysis [7] Pandian M, 2014, Land use and land cover changes detection using remote sensing and GIS in parts of Coibatore and Tiruppur districts, Tamil Nadu, India, Land use and land cover changes detection using remote sensing and GIS in parts of Coibatore and Tiruppur districts, Tamil Nadu, India [8] Robert A., Schowengerdt, 2007, Remote Sensing: Models andMethods for Image Processing, 3rd Edition, Oxford University, UK 49 [9] Selçuk Reis, 2008, Analyzing Land Use/Land Cover Changes Using Remote Sensing and GIS in Rize, North-East Turkey, Sensors ISSN 1424-8220 [10] Singh Bijender, 2014, Land Use / Land Cover Change of Delhi: A Study using Remote Sensing and GIS Techniques, International Research Journal of Earth Sciences [11] Tayyebi et al., 2008, Monitoring land use change by multi-temporal landsat remote sensing imager, University of Tehran, Iran [12] Trung Le Van, 2010, Remote Sensing, Publisher Vietnam National University, Ho Chi Minh city [13] Van Tran Thi; 2006; Application of Thermal remote sensing on increasing of urban surface temperature with distribution of land cover types in Ho Chi Minh city; Science & Technology Development, Enviroment &Resources, Volume – 2006 [14] Vimla Singh, 2012, Land Use Mapping Using Remote Sensing & GIS Techniques in Naina - Gorma Basin, Part of Rewa District, M.P, India, ISSN 22502459 [15] Vinh Tran Quoc, 2010, Using Satellite Data for Mapping Land Cover Factor (C) in Soil Erosion Research in Tam Nong District Phu Tho Province, Journal of Science and Development 2010: Volume 8, issue 6: 983 – 988 [16] Xiaoning Gong, Lars Gunnar Marklund, Sachiko Tsuji, 2009, Land Use Classification, FAO 50 Internet resource [17] Do Minh Phuong, Le Thi Kim Thoa, Le Duc Hoang, 2014, Remote Sensing application, building land cover map in Nghe An province http://www.ngheandost.gov.vn/ [18] Le Xuan Canh, Tran Anh Tuan, Remote Sensing & GIS application in management of forest resources, research result in Dak Lak http://lamdongdost.gov.vn/ [19] Satellite image http://earthexplorer.usgs.gov/ [20] Thach Nguyen Ngoc , 2012, Course of GIS http://www.ebook.edu.vn/ 51