38 11 : 1359 M I ( A B) = taila 熿 tailb taila B taila headb 燄 A tailb A B A headb 燀 heada tailb heada B heada head B 燅 (1) ( 1) HBT 2 HBT HBT 2 9 =512

Size: px

Start display at page:

Download "38 11 : 1359 M I ( A B) = taila 熿 tailb taila B taila headb 燄 A tailb A B A headb 燀 heada tailb heada B heada head B 燅 (1) ( 1) HBT 2 HBT HBT 2 9 =512"

其劳穆
5 years ago
Views:

1 GeomaticsandInformationScienceofWuhanUniversity Vol.38No.11 Nov.2013 : (2013) :A ( ) ( ) : 针对二维欧氏空间内有向线的特点将有向线对象区分为起点内部和边界建立了有向线对象之间拓扑关系与方向关系集成表达的 DLR 模型并给出了一个具体应用实例与现有模型进行对比分析发现 DLR 模型不仅能描述二维空间中有向线对象之间的方向关系还能够有效地区分相离内部相交相切部分覆盖等拓扑关系从而提高了有向线对象之间空间关系的区分能力 : 有向线 ; 拓扑关系 ; 方向关系 :P208 1 [1] [2] [3] RCC [4] [5] [6] [7] [8] [9] [12] Renz + [10] (DIA) [13] 26 JEPD DIA GPS [11] 1.2 Kura- ta [14] 9 X (tail) (head) tailx headx X HBT (head-body-tail) : : : ( ); (2012zzts008)

2 38 11 : 1359 M I ( A B) = taila 熿 tailb taila B taila headb 燄 A tailb A B A headb 燀 heada tailb heada B heada head B 燅 (1) ( 1) HBT 2 HBT HBT 2 9 = 熿燄燀燅 1 HBT Fig.1 BasicTpologicalRelationsDefinedbytheHBT Model 3 HBT 熿燄 HBT 燀燅 2 HBT [15] 4 Fig.2 DisjointRelationsnotBeing DDLO DistinguishedbytheHBT Mode 3 HBT Fig.3 CrossRelationsnotBeingDistinguishedbytheHBT Model DRA 2 B s B e B A A s A e A B : A R 1s B A R 2e B BR 3s A BR 4e A (2) ( 4): (l); (r); [9] (b); (s); [16] DRA [17] (i); (e);

3 (f) R {lrse} 24 [18] DRA c ; R {lrbsief} 72 DRA f 4 1 B A Fig.4 DirectionalRelationsBetweenaDirected LineandaPoint R 1 R 3 {sei} (disjoint) 1(a) 3 R 1=i B A (di- 3.1 DLR(DirectedLineRelation) vergedby) 1(c) ( ) 4 B A R 1=e A B (folow) 1(d) 5 B A R 3=s A B (precede) 1(e) 6 B A R 3=i (mergedby) DLR 1(f) : DLR ( A B) = ( AR 1 tailbar 2B AR 3 head B) (3) DLR(AB) A B 2) B A ; A ;B ;R 1 R 2 R 3 B B A A A : s i e R 2 = ( Ar 0B 0 Ar 1B 1 Ar nb n ) (4) (fl) (f) (fr) (l) (r) (bl) (b) (br) R i {sie 5 Fig.5 ReferenceSystemofDirectedLine SpatialRelation Model 2 B A R 1=s (split) 1(b) 3 B A 7 B A R 3= e (meet) 1 (g) B A i R 2 r j B j j+1 A r j {l ir}0 j n r j r flffrlrblbbr}1 i 3 5 r j j=l r j=r (cross) ( 6(a)) (touch)( 6(b)); n (n 1) (overlap)( 6(c)) 1) B A DLR 10 JEPD B A R 2 1 i B A DLR 7(a) DLR

4 38 11 : DLR Fig.6 Body-BodyIntersectionsBetweenDirectedLines (AB)= (ile) di- 2) DDLO vergedby-meet; 7(b) DLR(AB)=(s(r l)i) split-cross-merged- By; 7(c) DLR(AB)=(fr(rlil) bl) cross-overlap 1 DLR Tab.1 BasicTopologicalRelationsBetween DirectedLinesCorrespondingtoDLR 2 DLR : DLR DLR(AB)= (rfrfr) 2(a) ; R1R3 {sei} disjoint R1=s split DLR(AB)= (ffll) 2(b) ;DLR R1=i divergedby (AB)= (brbbl) 2(c) R2 i R1=e R3=s R3=i R3=e r j=i folow precede mergedby meet overlap i R2 r j=r j+1 touch r j=r j+1 cross 7 DLR(AB)= (br(rr)fr) Fig.7 ExamplesofSpatialRelationsBetween touch 3(c) ; DirectedLines DLR(AB)= (bl(lil)fl) overlap 3(d) ; 3.2 DLR(AB)= (bl(lrl)fl) DIA cross 3(e) HBT DDLO DRA DLR 2 2 DLR Tab.2 ComparisonofExisting ModelsandDLR DIA HBT DDLO DRA DLR HBT DDLO DRA DLR HBT cross/touch DLR 35 : DLR(AB)= (br(rl)fl) cross 3(a) ; DLR(AB)= (fl(lr)br) cross 3(b) ; 4 DLR ; 8(a) t 1 a 90 b ) DIA A B t 1

5 DLR(AB)= (flflfl) [6] Frank A U.QualitativeSpatialReasoning About t 2 Cardinal Directions[C]. Austrian Conference on 8(b) DLR(AB)= (ll ArtificialInteligenceAustralia1991 [7] r) PapadiasD.Relation-basedRepresentationofSpa- tialknowledge[d].athens:nationaltechnicalu- niversityofathens1994 [8] GoyalR.SimilarityAssessmentforCardinalDirec- [10]. 8 [J]. Fig.8 TrajectoriesofOceanVesselandTyphoon (9): [1] Egenhofer MFranzosa R.Point-set Topological SpatialRelationships[J].InternationalJournalof GeographicalInformation Systems19915(2): [16]Krieg-BrücknerBHuiShi.OrientationCalculiand Route Graphs:TowardsSemantic Representations forroutedescriptions[j].lecturenotesincom puterscience : [2] EgenhoferMSharmaJMarkD.A CriticalCom- [17]MoratzRLückeDMossakowskiT.A Condensed parisonofthe4-intesectionand9-intersection Mod- Semanticsfor QualitativeSpatialReasoning About elsforspatialrelations:formalanalysis[j].au- OrientedStraightLineSegments[J].ArtificialIn- tocarto199311:1-11 teligence (16/17): [3].GIS [18] MoratzRRenzJWolterD.QualitativeSpatial 4 [J]. ReasoningAboutLineSegments[C].The14thEu (11): ropean Conferenceon ArtificialInteligenceAm- [4] RandelD ACuiZCohn A G.A SpatialLogic sterdam2000 BasedonRegionsandConnection[C].The3rdIn- ternationalconferenceon Principlesof Knowledge RepresentationandReasoningSanFranciso1992 [5] Schneider MChen TaoViswanathan Getal. CardinalDirectionsBetweenComplexRegions [J]. ACM TransactionsonDatabaseSystems (2):1-45 tionsbetween ExtendedSpatialObjects[D].Oro- no:universityofmaine2000 [9] ZimmermannKFreksaC.QualitativeSpatialRea- soning Using Orientation Distance and Path Knowledge[J].AppliedInteligence19966(1): [11]ZhengYuZhouXiaofang.ComputingwithSpatial Trajectories[M].NY:Springer-Verlag [12]AlenJ.Maintaining Knowledge About Temporal Intervals[J].CommunicationsoftheACM (11): DLR DLR [13]RenzJ.ASpatialOdysseyoftheIntervalAlgebra: DirectedIntervals[C].The17thInternationalJoint ConferenceonArtificialInteligenceSeatle2001 [14]KurataYEgenhofer M.The Head-Body-TailIn- tersectionsforspatial Relations Between Directed DLR LineSegments[J].LectureNotesinComputerSci- DLR ence : [15]. [J]. ( )200939(5): : wj8211@126.com

6 38 11 : 1363 AnIntegrated ModeltoRepresentTopologicalRelationand DirectionalRelationBetweenDirectedLineObjects! WU Jing 1 DENG Min 12 LIU Huimin 1 (1 DepartmentofGeo-InformaticsCentralSouthUniversity932LushanSouthRoadChangsha410083China) (2 ResearchCenterofFoundationGeographicInformationEngineeringHunanProvince 693Shaoshan MiddleRoadChangsha410083China) Abstract:Duetotheincreasingapplicationdemandstheissuesonspatio-temporalcharacter- isticofmovingobjectshaveatractedawiderangeofatentioninthefieldsofcomputersci- enceandgeographicinformationscience.thespatialrelationsbetweentwodirectedlineob- jectsasoneofthemaintopicsaboutthetrajectoryofmovingobjectscontributesmuchto researchonspatio-temporaldatamodel.accordingtothecharacteristicsofdirectedlineob- jectsineuclideanspaceanintegrated modelofspatialrelationsbetweendirectedlineob- jectsnameddlrisproposedwhichdividesthedirectedlineobjectintoheadbodyand tail.themodelcannotonlydescribethedirectionalrelationsbetweendirectedlineobjects butalsodistinguishthetopologicalrelationsbetweenthemsuchasdisjointcrosstouch andoverlap.thatistosaythemodelproposedinthispaperisabletoenhancetheidentifi- cationcapabilityofspatialrelationsbetweendirectedlineobjects.anexampleoftyphoone- vasionofoceanvesselisgiventodemonstratetheapplicationofdlr. Keywords:directedlineobjects;topologicalrelation;directionalrelation Aboutthefirstauthor:WUJinglecturerPhDcandidatespecializedingeographicinformationmodelingtheoriesandmethodsofspatial relationships. wj8211@ 126.com 檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪檪 ( 上接第 1352 页 ) DiscriminantSpaceMethodforScalingLandCoverInformation YOU Jiong 1 ZHANG Jingxiong 2 (1 ChineseAcademyofAgriculturalEngineering41 MaizidianStreetBeijing100125China) (2 SchoolofRemoteSensingandInformationEngineeringWuhanUniversity129LuoyuRoadWuhan430079China) Abstract:Thispaperexploresascalingmethodforlandcoverinformationbasedondiscrimi- nantspacemodels.toobtainlandcoverclassificationaccordingtodiscriminantspacemodels atdiferentscalesthescalingofdiscriminantvectorswererealizedandthendiscriminant spacemodelscomposedofthesetransformeddiscriminantvectorswerefound.twoclassifi- cation methodsincludingthemaximumlikelihoodclassifierandsvm methodwereused.the resultsachievedwiththeproposedmethodwerecomparedwiththoseofconventionalscaling methodsonthebasisofmajoritytheoryandtheruleofcentriccelthusvalidatingtheefective- nessandtheavailabilityoftheproposed method.experimentalresultsshowthattheproposed methodhadbeterresultsandasmalerinformationlossthantheconventionalscalingmethods. Keywords:landcover;scaling;discriminantspacemodels;scaleefects;informationloss Aboutthefirstauthor:YOUJiongPhDspecializedinagriculturalremotesensinguncertaintyanalysisetc. rsjone@whu.edu.cn

39 3 : ( )TR μ ( A B; ) - ( )TR σ O A B= U F O (O UO F)U O O ( ) F O TR φ TR φ (τ)=τ Q O U F : O = U F Q : =U F Q O 1= ( )τ Q O-UO 2=O-FO 12=O-(

39 3 : ( )TR μ ( A B; ) - ( )TR σ O A B= U F O (O UO F)U O O ( ) F O TR φ TR φ (τ)=τ Q O U F : O = U F Q : =U F Q O 1= ( )τ Q O-UO 2=O-FO 12=O-( 39 3 2014 3 GeomaticsandInformationScienceofWuhanUniversity Vol.39No.3 March2014 DOI:10.13203/j.whugis20120732 :1671-8860(2014)03-0322-05 1 2 3 1 1 276000 2 518040 3 430079 : 分析了四维地籍的概念与需求提出了时空宗地的概念探讨了时空宗地的演化过程