Welcome to Journal of University of Chinese Academy of Sciences,Today is

Journal of University of Chinese Academy of Sciences ›› 2008, Vol. 25 ›› Issue (2): 257-263.DOI: 10.7523/j.issn.2095-6134.2008.2.018

Previous Articles    

ApplicationofBayesiantheoryinmultibandSARimagesfusion forclassification

YUEJin 1,2, YANGRu Liang, HUANRuoHong 1,2   

  1. 1InstituteofElectronics,ChineseAcademyofSciences,Beijing100080,China; 2GraduateUniversityoftheChineseAcademyofSciences,Beijing100049,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2008-03-15

Abstract: BayesiantheoryisappliedtoclassificationofmultibandSARimages.Thefamiliarfusion
methodslikeproduct,averagingandmedianrulesarediscussed,andonthebaseofBayesianaveragingmethod, threemodifiedrulesareproposedaccordingtotherelationshipbetweenclassificationaccuracyanddistancefactorof
SARimage.Experimentalresultsshowthatmultibandfusioncancombinethepreponderanceandcomplementary informationofeachbandandachieveclassificationresultswhichcouldn’tbegotinsingleband.Themodified methodsdecreasetheeffectofthewrongclassificationinformationthroughweightingandaccordinglyimprovethe classificationaccuracy.

Key words: SARimage, Bayesiantheory, multibandfusion, classification

CLC Number: