基于图象的数据挖掘方法研究

《基于图象的数据挖掘方法研究》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、西北工业大学硕士学位论文基于图象的数据挖掘方法研究姓名：段锋申请学位级别：硕士专业：计算机应用技术指导教师：张艳宁20030301基于匿象的数据挖掘方法研究西北工业大学硕士学位论文ABRSTRACTThedevelopingtrendoftheDataMiningisdiverse，whichtheflexibiiityofthetechnicofthedataminingisknownclearlY．Sincetheimprovementoftheeraoftheinformation．moretechniciswidelvinter

2、infiltratedwhichinvolvedthedatamining．Inthispaper，thetechnicofthedataminingisusedinthefieldoftheimageprocessing．Lotsoftheresearchingworkhasbeendone，whichcombinethetechnicofthedataminingtothetechnicoftheimageprocessing，tosolvetheprobleminremotesenseimageprocessing．Themode

3、lofthedataminingbasedonimagehasbeenbuiltandeverystepinthemodelhasbeenanalyzedtheoreticallyandappliedinreality．Themaincontentsofthisthesisaresummarizedasf01low：(1)ThecourseofthedataminingdevelopmentisreviewedandtherolewhichthedataminingplayedinKDDisdefinite．Thecomponentsh

4、asbeenclassifiedaccordingtothetaskofthedatamining．andineachcomponentthetechnichasbeensummarized．ThemodelwhichapplYtoimagedataminingisproposedbasedonthesecomponents．(2)Theimagedatareductionisveryimportantinremotesenseimageprocessing，twomethodsindatareductionproposedinthis

5、paper，whichoneisbasedonwavelettheoryandtheotherisbasedonclusteringanalysis．Themethodbasedonwavelettheoryisdiscussedintherespectoftherelationbetweenthewaveletfilterandthelinearf订ter．The9／7waveletliftingalgorithm．whichistheBiorthogonalwavelethasthelinearphasic，isappliedint

6、hereductionandhasagoodresult．Themethodbasedonclusteringanalysisputtheimageintosubimagebythequartertreemethodaccordingtothelocalcharacteristic，andinthesubimagefinishingclustering．Basedonthecenterofthesubimageclusteringwedotheclusteringtogettheclusteringonthewholeimage．(3)

7、Thedistributionandthedirectionintextureisdiscussed．Thetexturesegmentationmethodwhichbasedontheself—adaptedtexturewindowsselectedisproposed．TheGaborfilterisusedtoextractthecharacteristics．Inthismethod．thejetwhichisakindofthemulti-scaledanalysismethodsisused．andgetagoodres

8、ult．GMRFisakindofsegmentationmethodswhichcombinegraycharacteristieandtexturecharacteristicasitscharacte