高分辨率遥感影像信息提取与目标识别技术研究

,,,　,　(,100101)【　】由于高空间分辨率遥感影像海量数据、复杂细节和尺度依赖的特点决定了高分辨率遥感影像处理的技术难点。在总结以往高分辨率影像(航空影像)信息提取技术的主要难点和不足,从理论上和实践上分析了基于特征基元的高分辨率遥感影像处理与分析的意义,提出了基于特征基元的高分辨率遥感影像多尺度信息提取技术框架。最后对此框架进行总结与分析,指出了目前研究中仍存在的难点和今后的研究重点。【】高分辨率遥感;信息提取;目标识别;特征基元;影像分割;多尺度【】TP75　　　　【】A　　　　【】1009-2307(2005)03-0018-03:2004-08-12:(40101021);(CXIOG-D02-01)。1　　30,、、,、、、、/。、、,。SPOT-1,。,10m、5m、2m、1m、0.6m,“”。,、、、,,、。2　,。,。,、、,,,,。,、、(),,80%[1],,。TM()、SAR,,。()。、。,,Nagao,VirtuoZo。,:①,Nago;[2];,[3];[4]。②,,[5];[4]。③,,3DMOSAIC[6];DSM()GIS[7]。,,(2080),:,。,,、。,(/),、(//)(///)。[6]:①,。②“”,。,、。③,,。④,,,。。,“”,30320056　　　　　　　　　　　　　ScienceofSurveyingandMapping　　　　　　　　　　　　　Vol.30No.3Jun,。,。,,。:①;②;③Hough。Jong-hyeokjeongHough,,IKONOS[8],,BL[9,10]。,,,([11][12]),,。,:①,,,。②,,,、、,。,,,(),。③,,。④,,,,,。3　3.1　,,,,,,,:1),Blaschke,,,、、,。2),。,。Blaschke,,。3),,,。,,、、,,(per-pixel),,。,(per-fieldper-parcel)。,、“”。,(object-oriented),。,。4),。LoboAplin,,,,,;MauroHuangHuiping,per-fieldobject-oriented;O.Hellwich、;,Delphi2CreativeTechnologieseCognitionTM,object-oriented。,,,、。,,,,,。,。3.2　,(1)。。1,。,,,;,,,。,,,,,。:193　(),,1　(AOI)、、,,。,,。,、、、、。。,。,,,,、、,。。,,。,,,,(、、、)。,,GIS(、),,。,。,,,。,,。,。4　,:,,,,。,。,,,,。,,:,(),;,,(),;,,,,;,,,、、,、,,,。[1]　,,.———、[M].:,2004.[2]　,,.[J].,2000,25(3):238-242.[3]　,,.[J]..2002,28(8):20-22.[4]　,.[J]..2002,4.[5]　.[D][].:,1998.[6]　.[][D].:,2001.[7]　ClausB.Towardsfullyautomaticgenerationofcitymodels[A].In:InternationalArchivesofPho-togrammetryandRemoteSensing[C].Amsterdam:XIX.2000.[8]　Jong-HyeokJEONG,MasatakaTAKAGI.ExtractionOfBridgePositionsFromIkonosImagesForAccuracyControlofBridgeDatabase[EB/OL].[9]　,.[J].(),1995,29(3):275-281.[10]　,,.[J].,2002,(11):78-80.[11]　,,,.[J].,1997,20(12):1124-1128.[12]　,,,.[J]..2000,13(2):214-217.::(1976-),,。E-mail:[e-mail]20　　　30ElevationerrorestimationofsatellitephotogrammetrywithoutgroundcontrolpointAbstract:Thispaperdeduceselevationerrorestimatingformulaeofthesatellitephotogrammetry,whichisconcernedwithelementsofexteriororientationthatmeasuredbyspacebornesensors.Accordingtotheattitudestabilityoftwo-linearrayCCDpushbroomimageat10-6o/s,orlowerthan10-6o/sandtheattitudestabilityofsocalledLMCCD(line-matrixCCDcamara)pushbroomimagelowerthan10-6o/s,theauthorsdeduceandcalculatefordifferentcasesrespectively,andpresentanexampleforeachestimatingformulafortheapplicationofdesigningsatellitephotogrammetryengineeringsystemwithoutgroundcontrolpoint.Keywords:satellitephotogrammetry;errorestimation;CCDpush-broomimageWANGRen-xiang,LIJing,WANGXin-yi,YANGJun-fengResearchonspatial-temporaldatamodelinGISAbstract:StateandeventisapairofbasicconceptioninTemporalGeographicalInformationSystems(TGIS).Nowmostofspatio-temporaldatamodelsdealwithbothconceptionsseparately,andcannotcombinethemeritsofbothmodels.Inthispaper,anewspatial-temporaldatamodelbasedonstate-eventispresented,andthequalificationsforconversionbetweenspatial-temporaldatamodelsareanalyzed.Keywords:state;event;spatio-temporaldatamodel;spatio-temporalrelationshipYINZhang-cai①②,LILin②(①SchoolofResourceandEnvironmentProject,WuhanUniversityTechnology,129LuoyuRoad,Wuhan,China430079;②SchoolofResourceandEnvironmentScience,WuhanUniversity,129LuoyuRoad,Wuhan,China430079)ResearchonautomaticupdatingofGISdataAbstract:ItisvitaltokeepGISdata“fresh”forapplicationanddevelopmentofGIS.ThispaperdescribesageneralprocessforautomaticdataupdatingofGIS,discussesthreebasicmethodsofchangedetection,andanalysesthecharacteristicofthesemethods.Finally,theoutlookarepresentedforfurtherstudies.Keywords:GIS;changedetection;orthoimage;vector;rasterFANDa-zhao,ZHANGyong-sheng,LEIRong,GUOHai-tao(InstituteofSurveyingandMapping,InformationEngineeringUniversity,Zhengzhou450052)ResearchoninformationextractionandtargetrecognitionfromhighresolutionremotesensingimageAbstract:Highresolutionremotesensingimagehassomecharacteristicssuchastremendousdata,complexfeaturedetailsanddependenceonscale,sosomenewtechnicaldifficultyisfacedtoresearchersonremotesensing.Basedonsummarizingprevioustechniqueofinformationextractionfromhighresolutionimage(aerialimage),thispaperanalysesthesuperiorityofremotesensingimageprocessandanalysisbasedonparcelunitintheoryandinpractice.Furthermore,amulti-scaletechniqueforinformationextractionandtargetrecognitionfromhighresolutionremotesensingimagebasedonfeatureunitisbroughtforth.Intheend,thispaperexplainsthetechniqueframeandpointsouttheactualresearchdifficultiesandemphasisinthefuture.Keywords:highresolutionremotesensing;informationextraction;targetrecognition;parcelunit;imagesegmentation;multi-scaleMINGDong-ping,LUOJian-cheng,SHENZhan-feng,WANGMinM,SEHNGHao(TheStateKeyLaboratoryofResourcesandEnvironmentalInformationSystem,IGSNRR,CAS,Beijing100101)FractalattenuationanalysisofcartographicobjectoncartographygeneralizationAbstract:Onthecartographygeneralization,theattribute,structuralandshapeofthecartographyobjectwouldbechangedwithvariationofthescalesize,thereby,itresultsinthedecreaseofinformationascalledfractalattenuation.Thispaperusesfractalgeometryasatoolandfractaldimensionasap