遙感圖像中建筑物自動識別與標(biāo)繪方法研究.pdf

1、分類號!旦竺!：壘UDC密級編號中國科學(xué)院研究生院碩士學(xué)位論文y987841李海月指導(dǎo)教師王塞墮堡窺基生墾型堂瞳電壬堂堡宜壓申請學(xué)位級別亟學(xué)科專業(yè)名稱焦墨皇籃!量處堡論文提交日期2QQ魚璽量旦論文答辯日期至QQ魚生量目培養(yǎng)單位蟲墾型堂院皇王堂堡窒壓學(xué)位授予單位生國疊堂暄殛筮生隍一入。一TheTechniqueforAutomaticBuildingRecognitionandMappinginRemoteSensingImagesLil

2、。laiyuefSignalandInformationProcessing)DirectedbyProfWangHongqiAutomaticTargetRecognition(ATR)isacoretechnologyinthefieldofremotesensingimageprocessingRecentcommercialsatellites，likeIKONOSandQuickBird，providealargequanti

3、tyofhighresolutionremotesensingimagesTheATRtechnologyhasbecomemoreusefulandimportantThegoalofATRinremotesensingimagesistOrecognizetheregionsofinterest，suchaslbrests，lakes，crops，andbuildingsInordertOachievethisgoal，multip

4、letechniqueslikeimageprocessing，imagefeatureextraction，imagesegmentationareusedAfteratargetisrecognized，flmappingprocedureisneededtOacquiretheregularoutlineofthistargetAutomatictargetrecognitionandmappingtechniquesarewid

5、elyusedinenvironmentmonitoring，militarydetectiveagriculturalestimation，anddisastercontr01TheywillbethekeytechniquesoffutureGIS，digitalglobe，digitalcitysystemTheresearchesofthispapertbcusonautomaticbuildingrecognitionandm

6、appingtechniquesinurbanremotesensingimagesByconcludingandsummarizingformermethods，thispaperproposesfinapplicationframeworklbrautomaticbuildingrecognitionandmapping7rhcrearethreestepsinthisframework：candidatebuildingareas

7、earching，buildingtargetverification，andbuildingmappingInthecandidatebuildingareasearchingstep，animagesegmentationmethodbasedongreylevelCO—occurrencematrix(GLCM)isadoptedtogetcandidatebuildingareasinanurbanimageInthebuild

8、ingtargetverificationstep，amethodbasedonthedmninanllinesegmentisproposedThedominantlinesegment(thelongestlinesegment)ofeverycandidateareaisextractedbyaprocedureusingregiongrowing，edgedetection，andHoughtransformThen，there