建設(shè)用地及其變化信息遙感監(jiān)測方法研究.pdf

1、分類號：塑3DC：5269學(xué)號：鰻3籃iQ蟛密綴；玉建設(shè)用地及其變化信息遙感監(jiān)測方法研究馬雪梅指導(dǎo)教師姓名：毯蕉嫠麴援趣篷太堂盔巍湮丕縫堂院直壺直西鏖蹺!墨中請學(xué)位級別：蘭堂亟專業(yè)名稱：攝毖麴熏蔓遙壁論文提交時間：2QQ生2旦論文答辯時間：2QQ曼生墨旦學(xué)位授予單位和日期：邐盥太堂量鯉魚生旦答辯委員會主席：黃塞掛論文評閱人：黃塞拄、塞塹2006年3月中國南京專敝大剛每學(xué)～，■■^可頁弘頌AbstractTheurbanizationne

2、arlyreflectsthedevelopmentandprogressofallarea，markingcultureandcivilizationforoneeraWimthechangeoftheurbansystemtheurbanproblemisbecomingmorecomplicatedandisurgenttobesolvedStudyingthechangeofspaceexpandingandecologJica

3、lenvironmentofthecityisanimportwaytofindandresolvetheurbanproblemBuiltuplandexpansionisuppermostsymboloftheurbanspaceexpansionThepaperanalysesimageryfeatureofbuilt—uplandbasedon3StechnologiesandNanjingcityimageryin2002an

4、dthenclassifiesthebuiltuplandinformationbyusingdataminingtechnologytoclassificationoflandusingAsthesalnetimedataminingtechnologywasusedtodetectionofbuiltuplandchangeinformationAndsuperpositionanalysiswasappliedinresultof

5、changeinformationdetectedtocheckresultandrefinetheclassificationAndtheadaptabilityofthedetectionmethodWItSprovedThecoreofthestudyandconclusionsareasfollows：1ResearchontheautomaticextractionofbuildinginformationAtfirst，th

6、echaracteristicsofbuiltuplandand廿leirfeatureinyearimageryWel“ganalyzeddeeplyAndthebuiltuplandoffeaturesetswasmadeAtthe88naetimetheregioninwhichobjectsweredispersedandcomplex，Wassegmentedintothreesubregionsinwhichobjectsh

7、adrepresentationbyusingregiondependentsegmentationtechnologyThenthedataminingtechnologyWasusedtoextractoptimumfeaturesubsetofbuiltuplandandcorrespondingknowledgerulesAfterth峨thedecisiontreealgorithmwasusedtoclassifythest

8、udyareaandconstructionlandusinginformationwasextractedAccuracyofclassificationofdecisiontreeWas847％andthefirstclassificationaccuracyis916％ItWASshowedthatitcouldworkwellintheautomaticextractionofbuiltuplandandhadthevalueo

9、fusingThetechniquew越easy幻beemployedandthegeneratedknowledgerulesareobjective2Researchonthechangedetectionofbuiltuplandusing碭echaracteristicsofchangeofbuiltuplandandtheirfeatureinyearimagerywereanalyzeddeeplyAndtheconstru

10、ctionoffeaturesetsWasmadeThedataminingtechnologyinmachinestudywasusedtodetectthechangeofbuiltupland，andtheresultofstairoriginalclassificationwasgainedThentheresultofclassificationWasfoldanalyzedbythebuilt—uplandexffactio

11、nofformerage(forexamplethreetypesconstructionlandusingofyear2002)，andwasproofedbywhichthestairandsecondclassificafionwereobtmnedFinallythesat／lechangedetectiontechnologyflowWasusedtogetchangeinformationbetween1988and1994

12、111eresultoffoldanalysisandaccuracyoftheresultshowthatdataminingtechnologyusedtodetectthechangeWasfeasibleAtthesalnetimeontheconditionoflackgrounddata，usingtheresultofformerageclassificationcouldproofthechangedetectionre