Classification of wetland from TM imageries based on decision tree

The traditional method of application of remote sensing data for land cover mapping is the use of supervised classification and unsupervised classification. Decision tree, showing great advantages in remote sensing classification, is computationally fast, makes no statistical assumptions, and can handle data that are represented on different measurement scales. Decision tree classification has been successfully applied to many classification problems, but rarely applied to mapping of wetlands. In this study, decision tree was proposed to extract wetland from Landsat 5/Thematic Mapper (TM) imageries in a wide area of Yinchuan plain. Tasseled Cap (TC) transformation was used to identity the different wetland types and normalized difference vegetation index (NDVI) was computed to distinguish paddy wetland and lake wetland. Results from this analysis show that the decision tree has an outstanding performance compared with the supervised classification in maximum likelihood method. The overall accuracy of supervised classification is 64.60%, while that of decision tree classification was 83.80%. Besides, it appears that a decision tree combinations different useful knowledge is an effective and promising classification method.

[1]  Paul M. Mather,et al.  An assessment of the effectiveness of decision tree methods for land cover classification , 2003 .

[2]  Eric P. Crist,et al.  A Physically-Based Transformation of Thematic Mapper Data---The TM Tasseled Cap , 1984, IEEE Transactions on Geoscience and Remote Sensing.

[3]  E Brown de Colstoun,et al.  National Park vegetation mapping using multitemporal Landsat 7 data and a decision tree classifier , 2003 .

[4]  Hasan Roosta,et al.  Comparison between Sub-pixel Classifications of MODIS images: Linear Mixture Model and Neural Network Model , 2008 .

[5]  Sholom M. Weiss,et al.  Computer Systems That Learn , 1990 .

[6]  Chao Zhang,et al.  Comparison of pixel-based and object-oriented knowledge-based classification methods using SPOT5 imagery , 2009 .

[7]  John R. Jensen,et al.  Introductory Digital Image Processing: A Remote Sensing Perspective , 1986 .

[8]  Lei Zhang,et al.  Evolutionary flexible neural networks for intrusion detection system , 2006 .

[9]  C. Woodcock,et al.  An assessment of several linear change detection techniques for mapping forest mortality using multitemporal landsat TM data , 1996 .

[10]  P. Utgoff,et al.  Multivariate Versus Univariate Decision Trees , 1992 .

[11]  Suming Jin,et al.  Comparison of time series tasseled cap wetness and the normalized difference moisture index in detecting forest disturbances , 2005 .

[12]  Ian H. Witten,et al.  Data mining: practical machine learning tools and techniques, 3rd Edition , 1999 .

[13]  Michael F. Goodchild,et al.  Using NDVI to define thermal south in several mountainous landscapes of California , 2009, Comput. Geosci..

[14]  W. Cohen,et al.  Estimating structural attributes of Douglas-fir/western hemlock forest stands from Landsat and SPOT imagery , 1992 .

[15]  E. Small,et al.  The impact of soil reflectance on the quantification of the green vegetation fraction from NDVI , 2005 .

[16]  David A. Landgrebe,et al.  A survey of decision tree classifier methodology , 1991, IEEE Trans. Syst. Man Cybern..

[17]  R. Congalton,et al.  Accuracy assessment: a user's perspective , 1986 .

[18]  R. J. Kauth,et al.  The Tasseled Cap de-mystified , 1986 .

[19]  Wei-Yin Loh,et al.  Classification and regression trees , 2011, WIREs Data Mining Knowl. Discov..

[20]  P. Kokol,et al.  End user friendly data mining with decision trees: a reality or a wish? , 2007 .

[21]  R. DeFries,et al.  Classification trees: an alternative to traditional land cover classifiers , 1996 .

[22]  C. Brodley,et al.  Decision tree classification of land cover from remotely sensed data , 1997 .

[23]  N. Coops,et al.  Extracting urban vegetation characteristics using spectral mixture analysis and decision tree classifications. , 2009 .

[24]  Peter Norvig,et al.  Artificial Intelligence: A Modern Approach , 1995 .

[25]  R. Congalton.  A Quantitative Method to Test for Consistency and Correctness in Photointerpretation , 1983 .

[26]  J. Dave,et al.  Influence of illumination and viewing geometry and atmospheric composition on the “tasseled cap” transformation of landsat MSS data , 1981 .

[27]  Ian H. Witten,et al.  Data mining - practical machine learning tools and techniques, Second Edition , 2005, The Morgan Kaufmann series in data management systems.

[28]  R. Smith,et al.  An approach for assessing wetland functions using hydrogeomorphic classification, reference wetlands, and functional indices; [computer file] /; by R. Daniel Smith ... [et al.] ; prepared for U.S. Army Corps of Engineers. , 1995 .

[29]  Ian Witten,et al.  Data Mining , 2000 .

[30]  E. Crist.  A TM Tasseled Cap equivalent transformation for reflectance factor data , 1985 .

[31]  Xiang Zhu,et al.  A hierarchical object oriented method for land cover classification of SPOT 5 imagery , 2009 .