Abstract：K-means clustering algorithm is a kind of hard classification based on the Euclidean distance, with each data point assigned to a single cluster. Due to the uncertainty and mixed pixels in remote sensing image,it is difficult for the traditional K-means clustering algorithm to obtain satisfactory classification results. To overcome this drawback,the authors applied the SPA(set pair analysis)theory to the clustering algorithm of remote sensing image. The IDC(identical discrepancy contrary)connection degree model,which can descript unitarily the identity,discrepancy and opposition,was employed to improve K-means clustering algorithm. The improved algorithm has overcome the limitation of K-means clustering algorithm to certain extent. Clustering analysis experiments of Landsat TM image show that the improved K-means clustering algorithm is superior to K-means in classification accuracy of ground cover class components of mixed pixels.
Zhao Y S,Chen D M,Yang L M,et al.Principles and Methods of Remote Sensing Applications[M].Beijing:Science Press,2003(in Chinese).
[2] Huang Z X.Extensions to the K-means Algorithm for Clustering Large Data Sets with Categorical Values[J].Data Ming and Knowledge Discovery,1998,2(3):283-297.
[3] Kaufman L,Rousseeuw P J.Finding Groups in Data:An Introduction to Cluster Analysis[M].Beijing:Wiley Online Library,1990.
[4] Hansen P,Jaumard B.Cluster Analysis and Mathematical Programming[J].Math Program,1997,79(1/3):191-215.
Deng X J,Wang Y P,Peng H L.The Clustering of High Resolution of Remote Sensing Imagery[J].Journal of Electrics and Information Technology,2003,25(8):1073-1080(in Chinese with English Abstract).
Zhong Y F,Zhang L P.Initialization Methods for Remote Sensing Image Clustering Using K-means Algorithm[J].Journal of System Engineering and Electronics,2010,32(9):2009-2014(in Chinese with English Abstract).
Liu X F,He B B,Li X W.Classification for Beijing-1 Micro-satellite’s Multispectral Image Based on Semi-supervised Kernel FCM Algorithm[J].Acta Geodaetica et Cartographica Sinica,2011,40(3):301-306(in Chinese with English Abstract).
Hasi B G,Ma J W,Li Q Q,et al.Improved Fuzzy C-mean Classifier and Comparison Study of Its Clustering Results of Satellite Remotely Sensed Data[J].Computer Engineering,2004,30(11):14-15(in Chinese with English Abstract).
[9] Dunn J C.A Fuzzy Relative of the ISODATA Process and Its Use in Detecting Compact Well Separated Cluster[J].Cybernetics and Systems,1974,3:32-57.
[10] Bezdek J C.Pattern Recognition with Fuzzy Objective Function Algorithms[M].New York:Plenum Press,1981.
Zhao K Q.The Application of Uncertainty Systems Theory of Set pair Analysis (SPA) in the Artificial Intelligence[J].CAAI Transactions on Intelligent Systems,2006,1(2):16-25(in Chinese with English Abstract).
[12] 赵克勤.集对分析及其初步应用[M].杭州:浙江科学技术出版社,2000.
Zhao K Q.Set Pair Analysis and Its Preliminary Application[M].Hang Zhou:Zhejiang Science and Technology Press,2000(in Chinese).
[13] Ball G H.ISODATA,a Novel Method of Data Analysis and Pattern Classification[R].Menlo Park:DTIC Document,1965.
[14] Lloyd S.Least Squares Quantization in PCM[J].IEEE Transactions on Information Theory,1982,28(2):129-137.
[15] MacQueen J.Some Methods for Classification and Analysis of Multivariate Observations[C] //In:Fifth Berkeley Symosium on Mathematics.Statistics and Probability.California:University of California Press,1967:281-297.
[16] Steinhaus H.Sur la Division Des Corp Materiels en Parties[J].Bull Acad Polon Sci,1956,4(1):801-804.
Zhao K Q.The Classification,Measurement and Applications Based on Set Pair Analysis[J].Science,Technology and Dialectics,1994,11(2):26-30(in Chinese with English Abstract).
Standardization Administration and General Administration of Quality Supervision,Inspection and Quarantine of the People’s Republic of China.GB/T 21010-2007 The Classification of Current Land Use[S].Beijing:China Standards Publishing House,2007(in Chinese).
2012, Vol. 24 
