Full-Text Search:
Home|Journal Papers|About CNKI|User Service|FAQ|Contact Us|中文
《Wetland Science》 2011-03
Add to Favorite Get Latest Update

Remote Sensing Classification for Zhalong Wetlands Based on Support Vector Machine

ZHANG Ce,ZANG Shu-Ying,JIN Zhu,ZHANG Yu-Hong(College of Geographical Sciences,Harbin Normal University,Harbin 150025,Heilongjiang,P.R.China)  
Wetlands are integral parts of the global ecosystem as they can prevent or reduce the severity of floods,feed ground water,and provide unique habitats for flora and fauna.Wetland remote sensing classification is one of the most important means to realize wetland dynamic monitoring,management and utilization.But spectral uncertainty or vagueness caused by spectral confusion between-class and spectral variation within-class remains a challenge to the wetland remote sensing classification.Exploring small sample size,high accuracy classification methods are hence necessary,due to the complex aquatic-terrestrial environments of wetlands,which are relatively difficult to acquire training samples by field survey.Traditional remote sensing classification methods usually require large training data sets,for instance,maximum likelihood classification,a widely used classical supervised classification method in remote sensing.The method of maximum likelihood classification assumes that the training statistics for each class have a normal distribution.Training of a maximum likelihood classification aims at a complete description of each class.To achieve this goal,the training set should be sufficiently large.Furthermore,the training data instances should be typical and representative of the classes in order to derive appropriate training statistics on which the classification can be based.It is hard to find such training samples in wetland remote sensing classification due to the wetland's complexity and ecological vulnerability.New approaches which need small training data sets and could deal with high feature dimensions remain to be investigated.In this paper,wetland remote sensing classification using support vector machine was explored by using Zhalong National Nature Reserve.The support vector machine is a new generation of supervised learning system that is based on the principle of statistical learning theory,which had gained wide acceptance in the remote sensing community for its solid statistical foundation and excellent empirical performance.One of the main advantages of support vector machine regarding remote sensing classification is that the approach only considers samples that are close to the class boundaries,the so-called support vectors.Using this,support vector machine works satisfactorily even with small training sets.Zhalong National Nature Reserve is a typical fresh water wetland area dominated by reed vegetation type.As an important breeding habitat of Red Crown crane(a rare wetland bird),Zhalong National Nature Reserve was firstly listed as the wetlands of international importance by the Ramsar Convention in 1987.The scientific management and protection of Zhalong National Nature Reserve has been highlighted by the government.Primary analysis was made with respect to the impact of sample size and feature dimensionality on the classification accuracy of the support vector machine,in comparison with classical maximum likelihood classifier.Training samples were partitioned into five sample sizes per class 20,40,60,80 and 100,based on stratified random sampling strategy.Each group of sample size combined with low and high dimensional features were used for wetland remote sensing classification respectively.The low dimensional features only include 11 feature bands while high dimensional features consist of 44 feature bands:11 feature bands and spectral,textural and terrain features.The wetland remote sensing classification based on support vector machine and maximum likelihood classifier had been explored using five kinds of sample size with low dimensional feature and high dimensional feature respectively.Our results showed that support vector machine performed better than maximum likelihood classifier in general.The classification accuracy tends to increase with the sample size grows larger.The highest classification accuracy(overall accuracy up to 88.125%) was obtained by support vector machine with 100 training samples per class plus high dimensional features.The total wetland area classified in Zhalong National Nature Reserve was 90 307.17 ha,including 8 301.15 ha water body,and 33 063.57 ha water covering marsh and 48 942.45 ha without water covering marsh.With high dimensional feature,support vector machine successfully solved the problem of'dimension disaster' and obtained an ideal classification result by combining spectral,textural and terrain features simultaneously.The support vector machine could still achieve good accuracies even the training sample size is as small as 20 or 40.Whereas maximum likelihood classifier classifier could not be trained with such small sample sizes at all,which indicate that maximum likelihood classifier classifier has a great constraint on sample sizes,especially for high dimensional features.Considering of its less constraints on training sample sizes and of its higher capability in dealing with dimensional features,the support vector machine is thus considered to be better suitable to wetland remote sensing classification,as training samples are hard to be acquired by field survey.That is,the support vector machine is an effective tool in wetland remote sensing classification.Meanwhile,we found 9 532.8 ha Saline-alkaline land in Zhalong National Nature Reserve.That means that the ecological environment of Zhalong National Nature Reserve is particularly vulnerable.Actions of scientific protection and management are required for these valuable wetland resources.
【Fund】: 国家自然科学基金重点项目(41030743)资助
【CateGory Index】: TP751.1
Download(CAJ format) Download(PDF format)
CAJViewer7.0 supports all the CNKI file formats; AdobeReader only supports the PDF format.
【Citations】
Chinese Journal Full-text Database 5 Hits
1 LUO Jian cheng 1, WANG Qin min 2,MA Jiang hong 3, ZHOU Cheng hu 1, LEUNG Yee 4 (1. Institute of Geographical Sciences and Natural Resources Research, CAS, Beijing 100101, China; 2. Fuzhou University,Fuzhou 350002,China; 3. Xi'an Jiaotong Unive;The EM-based Maximum Likelihood Classifier for Remotely Sensed Data[J];Acta Geodaetica Et Cartographic Sinica;2002-03
2 ZHANG Rui1,2,MA Jianwen3(1.Institute of Remote Sensing Applications,CAS,Beijing 100101,China; 2.Graduate University of CAS,Beijing 100049,China;3.Center for Earth Observation and Digital Earth,CAS,Beijing 100190,China);State of the Art on Remotely Sensed Data Classification Based on Support Vector Machines[J];Advances in Earth Science;2009-05
3 YAO Yun-jun1,ZHANG Ze-xun1,2,QIN Qi-ming1,QIU Yun-feng1(1.Institute of Remote Sensing & GIS,Peking University,Beijing 100871,China;2.School of Information,Yanshan University,Qinhuangdao Hebei 066004,China);Study on wetland information extraction of remote sensing images based on support vector machine[J];Application Research of Computers;2008-04
4 <author> Wei-Hong, YANG Liu, ZHANG Zheng-Xiang(Key Laboratory of Wetland Ecology and Vegetation Restoration for National Environmental Protection, Northeast Normal University, Changchun Jilin 130024, China; Ecological Environment Research Center, Chinese Academy of Sciences, Beijing 100085, China);<title>thod of Wetland Classification Based on Landsat7 ETM + Image[J];Wetland Science;2004-03
5 ZHENG Xiao-Yan,ZHANG Yan-Hong,LIU Wan-Song (College of Geo-exploration Science and Technology,Jilin University,Changchun 130026,Jilin,P. R. China);Method on Cloud Shadow Removing from Water Bodies in the Process of Remote Sensing Extraction in Zhalong Wetlands[J];Wetland Science;2010-01
【Co-citations】
Chinese Journal Full-text Database 10 Hits
1 LIU Li-quan et al(Liaoning Normal University,Dalian,Liaoning 116029);Study on the Extraction Method of the Information of Wolong Lake Wetland[J];Journal of Anhui Agricultural Sciences;2010-04
2 ZENG Tao;DI Xue-ying;YU Hong-zhou;SHU Zhan;YANG Guang College of Forestry,Northeast Forestry University,Harbin,150040,P.R.China;Changes of landscape patterns in Xingkai Lake Nature Reserve,northeastern China[J];Journal of Beijing Forestry University;2010-04
3 CHEN Liang1,LIU Xi2,ZHANG Yuan1(1.State Key Laboratory of Hydrology-Water Resource and Hydraulic Engineering,Hehai University,Nanjing 210098,China;2.The Third Institute of Oceanography,SOA,Xiamen 361005,China);A study of image classification based on MLC combined with spectral angle[J];Engineering of Surveying and Mapping;2007-03
4 YANG Honglei1,PENG Junhuan1,LI Shuhui1,SHI Yun1,2 1.School of Land Science and Technology,China University of Geosciences,Beijing 100083,China;2.Department of Surveying Engineering,X i'an University of Science and Technology,X i'an 710054,China;Log-principal Component Transformation Based EM Algorithm for Remote Sensing Classification[J];Acta Geodaetica et Cartographica Sinica;2010-04
5 Han YueJiao,Wang ChongChang (School of geomatics Liaoning,Technology University,Fuxin 123000,China);Remote Sensing Image Classification Method and Discussion Based on the TM[J];Urban Geotechnical Investigation & Surveying;2009-06
6 ZHAO Hui1,2,ZHANG Hai-ying2,LI Na-na3,YANG Gui-fang1 (1.School of Earth Sciences and Resources,China University of Geosciences,Beijing 100083;2.Institute of Remote Sensing Application,State Key Laboratory of Remote Sensing Science,CAS,Beijing 100101;3.China University of Mining and Technology,Beijing 100083,China);RS-Based Research of Chinese Wetland:Status and Trend[J];Geography and Geo-Information Science;2010-02
7 LI Jianping1,3,ZHANG Bai1,ZHANG Ling2,WANG Zongming1,SONG Kaishan1(1.Northeast Institute of Geography and Agriculture Ecology,CAS,Changchun 130012,China; 2.Geo-information Engineering Institute of Jilin province,Changchun 130061,China; 3.Graduate University of Chinese Academy of Sciences,Beijing 100039,China);Current Status and Prospect of Researches on Wetland Monitoring Based on Remote Sensing[J];Progress in Geography;2007-01
8 HASI Ba-gan,MA Jian-wen, LI Qi-qing(The Institute of Remote Sensing Applications,CAS, Beijing 100101,China);STUDY ON ASTER DATA CLASSIFICATION USING SELF-ORGANIZING NEURAL NETWORK METHOD[J];Advance in Earth Sciences;2003-03
9 ZHAO Shu-he~(1,2)1.Department of Urban & Resources Sciences,Nanjing University,Nanjing 210093,China2.Institute of Remote Sensing & GIS,Peking University,Beijing 100871,China;Experimental study of methods for high spatial resolution remote sensing data processing[J];Earth Science Frontiers;2006-03
10 ZHOU Xu1,LIU Zhi-hui1,2,3, DAI Wei1,JU Chun-yan1,YI Li1,JIANG Hong1 (1.College of Environmental Science and Engineering,Sun Yat-Sen University,Guangzhou,Xinjiang 510275;2.International Center for Desert Research on Sustainable Development in Arid and Semi-arid Lands,Urumqi,Xinjiang 830046;3.Key Laboratory of Oasis Ecology of Xinjiang Uygur Autonomous Region,Xinjiang University,Urumqi,Xinjiang 830046,China);Analysis based on RS & GIS of the effect of land-use/cover change on vegetation carbon reserves in arid region——A case study on oases of Shanshan,Xinjiang[J];Agricultural Research in the Arid Areas;2007-06
China Proceedings of conference Full-text Database 1 Hits
1 Yang Wu,Chunhui Wang,Le Yu Institute of Spatial Information Technique,Zhejiang University,Hangzhou,China Dengrong Zhang Institute of Remote Sensing and Geoscience,Hangzhou Normal University,Hangzhou,China;Using MRF approach to wetland classification of high spatial resolution remote sensing imagery:A case study in Xixi Westland National Park,Hangzhou,China[A];[C];2010
【Secondary Citations】
Chinese Journal Full-text Database 10 Hits
1 YANG Cun jian 1,2 ,ZHOU Cheng hu 3 (1.Institute of Remote Sensing Applications, Beijing 100101; 2. Yunnan University, Kunming,650091; 3.LREIS, Chinese Academy of Sciences, Beijing 100101,China);Investigation on Classification of Remote Sensing Image on Basis of Knowledge[J];Geography and Territorial Research;2001-01
2 HE Yu-hua,XIE Jun-qi,LIU Shun-xi(China Land Surveying and Planning Institute,Beijing 100035,China);Image Characteristics Analysis and Application Accuracy Assessment of ALOS Data[J];Geography and Geo-Information Science;2008-02
3 Fang Hongliang (Institute of Geography, Chinese Academy of Sciences, Beijing\ 100101);A DISCUSSION ON TWO STRATEGIES APPLIED TO ESTIMATE RICE PLANTING AREA OF AN ADMINISTRATIVE DIVISION USING REMOTE SENSING TECHNIQUE[J];ACTA GEOGRAPHICA SINICA;1998-01
4 ZHONG Chunqi,ZENG Congsheng,LIU Zhengzheng (College of Geographical Science of Fujian Normal University,Subtropical Wetland Center of Fujian Normal University,Fuzhou 350007,China);Study on Terrestrial Water Information Identified Based on the Analysis of Spectral Signature and Ratio Index[J];Geo-Information Science;2008-05
5 YANG Yong xing (Northeast Institute of Geography and Agricultural Ecology, CAS, Changchun 130021, China);NEW KNOWLEDGE ON THE PROGRESS OF INTERNATIONAL WETLAND SCIENCE RESEARCH AND PRIORITY FIELD AND PROSPECT OF CHINESE WETLAND SCIENCE RESEARCH[J];Advance In Earth Sciences;2002-04
6 LI Shuang 1,2 , ZHANG Re-xun 1 (1.College of Environment and Planning, Henan University, Kaifeng 475001, China; 2.Department of Geography, Liaocheng Teachers' University, Liaocheng 252059, China);The Decision Tree Classification and Its Application in Land Cover[J];Areal Research and Development;2003-01
7 FAN Wei, XUN Shang-pei, WU Wen-yu (Key Lab. of Atmospheric Science and Satellite Remote Sensing, Anhui Meteorological Institute, Hefei 230031, China);Waterbody Identification Under Semitransparent Cloud in MODIS Image[J];Journal of Atmospheric and Environmental Optics;2007-01
8 CAO Kai~(1,3),JIANG Nan~3,LI Xin-guo~(2,3),L Heng~3 (1.Department of Urban and Resource,Nanjing University,Nanjing 210093,China;2.Graduate School of The Chinese Academy of Sciences,Beijing 100039,China;3.Nanjing Institute of Geography and Limnology,Chinese Academy of Sciences,Nanjing 210008,China);A MODEL FOR AUTOMATIC EXTRACTION OF WATER INFORMATION FORM THE CITY ZONE BASED ON SPOT-5 IMAGE[J];Remote Sensing For Land & Resources;2005-04
9 <author> Wei-Hong, YANG Liu, ZHANG Zheng-Xiang(Key Laboratory of Wetland Ecology and Vegetation Restoration for National Environmental Protection, Northeast Normal University, Changchun Jilin 130024, China; Ecological Environment Research Center, Chinese Academy of Sciences, Beijing 100085, China);<title>thod of Wetland Classification Based on Landsat7 ETM + Image[J];Wetland Science;2004-03
10 TONG Shou-Zheng 1,2,LU Xian-Guo2,SU Li-Ying3,JIANG Ming2,YAO Yun-Long2(1.College of Urban and Environmental Sciences,Northeast Normal University,Changchun 130024,Jilin,P.R.China;2.Key Laboratory of Wetland Ecology and Environment,Northeast Institute of Geography and Agroecology,Chinese Academy of Sciences,Changchun 130012,Jilin,P.R.China;3.International Crane Foundation World Center for the Study and Preservation of Cranes,Baraboo WI 53913,Wisconsin,U.S.A.);Changing Process and the Impact Factors of Wetland Ecosystem in Zhalong Wetland[J];Wetland Science;2008-02
©2006 Tsinghua Tongfang Knowledge Network Technology Co., Ltd.(Beijing)(TTKN) All rights reserved