Fig.1 Arterial-roadPolygonsClassification compactness=4π area (polygon) perimeter 2 (polygon) (1) [0,1], 2), (minimum boundaryrectan- gul

Size: px

Start display at page:

Download "Fig.1 Arterial-roadPolygonsClassification compactness=4π area (polygon) perimeter 2 (polygon) (1) [0,1], 2), (minimum boundaryrectan- gul"

蛟备裘
7 years ago
Views:

1 GeomaticsandInformationScienceofWuhanUniversity Vol.39No.3 March2014 DOI /j.whugis (2014) ,2,3 4 1,2 5 1,, ,, ,, ,, ,, 基于多边形的形态分析提出一种城市主干道提取方法首先根据点线数据生成多边形并计算多边形几何形态指标 ; 然后使用支持向量机集成各项指标对生成的多边形进行形态分类, 提取候选主干道多边形 ; 最后根据格式塔理论使用区域增长算法连接候选主干道多边形, 提取最终的道路网主干道实验表明, 通过本方法能够快速有效地提取道路网中的平行车道与道路属性数据中的高等级道路比较发现, 本文提取的主干道与道路网的建设等级趋于一致交通地理信息系统 ; 主干道 ; 道路网 ; 格式塔 ; 道路等级 P208! A,,, [1,2] 1, [3-8] 1.1, [9,10], ArcGIS, 1,, 3 [11], ;,,,,, 5,, 3 1),, (2012AA12A211,2012AA12A204); ( ); ( , ); ( ); ( ); ( ),, xuechen.luan@whu.edu.cn, hongchao@tongji.edu.cn

2 Fig.1 Arterial-roadPolygonsClassification compactness=4π area (polygon) perimeter 2 (polygon) (1) [0,1], 2), (minimum boundaryrectan- gular,mbr) paralelism = length (main direction) perimeter(polygon) ( 2) 1.2 (0,1) 5,,,,,,, (SVM) 3) 5 SVM elongation= length2 (centerline) area(polygon) (3) SVM, (0, ) - (x i,yi),i = 1,2,, l, x i,yi {-1,1} i (0,1] [13],SVM [12] Delaunay, 3 min ( 1 l w,b, ξ 2 wt w+c ξ ) (4) i i=1 2,Ⅰ Ⅲ yi(w T φ ( x i )+b) 1-ξ, i ξ i 0,C>0 Ⅰ Lagrange Ⅰ A P 1 ;Ⅱ P 1 P 2, Ⅲ O 3 P 1 f(x)=sgn{ l i=1 α* iyik(x i,x)+b * }(5) P 2 P 3,α * b *, K 3, (x i,x)=(x i ) T (x) x,, Delaunay (RBF), RBF 2 [12] Fig.2 TrianglesClassification, 3 Fig.3 ExampleofCentralLineExtraction,,

3 , SVM SVM libsvm [14] 2 SVM, [15] 4 Fig.4 ArrangementofPolygons SVM, P i P j, P i P j 4 B ij L boundary, P i P j L i W i L j W j Delaunay (OSM),, Angle ij Angle ib 45, 4(a), SVM L new =L i +L j -L boundary (6) Angle ij 45, 4 (b), L new =L i +W j +L j -L boundary (9), OSM,, W new = (area i +area j )/L new (10) 3), 30m, W new ; 1), P i P i, P j B ij 4), 2) Angle ij P i P j,angle ib B ij P i, Angle ij Angle ib P i, P j 3, 3 5,, 3 L new =L i +L j (7) 93.79% Angle ij 45, 89.44% 91.14% Angle ib 45, 6 4(c) L new =L i +W i +L j -L boundary (8), Angle ib 45, 4(d),,,

4 Fig.5 Arterial-roadExtractionResult (6) (,,. 6 [J]. Fig.6 StatisticalResultofComparisonBetweenOur,2012,37(6) ) MethodandConstructedRoadLevels [5] PortaS,CrucitiP,LatoraV.TheNetworkAnal-, ysisofurbanstreetsadualapproach[j].physi- ca AStatistical MechanicsandIts Applications, 2006,369(2) [6] JiangB,ClaramuntC.A StructuralApproachto the ModelGeneralizationofan UrbanStreet Net- work[j].geoinformatica,2002,8(2) [7] Tomko M, WinterS,ClaramuntC.Experiential HierarchiesofStreets[J].Computers,Environment and Urban Systems,2008,32(1)41-52 [8] DengYajuan,YangYunfeng,MaRongguo.High- [1] YangBisheng,SunLi.Adaptive Multi-scaleVisu- way Network Structure Characteristics Based on alizationsofroadnetworkfornavigation[j].geo- Complex Network Theory[J].China Journalof maticsandinformation Scienceof Wuhan Univer- Highway and Transport,2010,23(1)98-104( sity,2008,33(4) (,. [J].,2008,33(4)37-41) [2] LiDeren,LiQigquan,YangBisheng,etal.Tech- niquesofgis,gpsandrsforthedevelopmentof InteligentTransportation[J].Geomaticsand In- formation Scienceof Wuhan University,2008,33 (4) (,,,.3S [J]., 2008,33(4)37-41) [3] LiQingquan,ZengZhe,YangBisheng,etal.Be- tweennesscentralityanalysisforurbanroadnet- works[j].geomaticsandinformation Scienceof Wuhan University,2010,35(1)37-41(,, [J].,2010,35(1) 37-41) [4] Luan Xuechen, Yang Bisheng, Zhang Yunfei. StructuralHierarchy AnalysisofStreetsBasedon Complex Network Theory[J].Geomaticsand In- formation Scienceof Wuhan University,2012,37,,. [J].,2010,23(1)98-104) [9] LuanXuechen,YangBisheng.GeneratingStrokes from CityRoadNetworks[J].Geography and Geo- Information Science,2009,25(1)49-52(,. Stroke [J].,2009,25(1)49-52) [10]YangBS,LuanXC,LiQ Q.Generating Hierar- chicalstrokesfrom UrbanStreet NetworksBased onspatialpatern Recognition[J].International Journal of Geographical Information Science, 2011,25(12) [11]TouyaG.A RoadNetworkSelectionProcessBased,. on Data Enrichmentand Structure Detection[J].

5 TransactionsinGIS,2010,14 (5) [12]AiTinghua,GuoRenzhong.A ConstrainedDelau- naypartitioningof ArealObjectstoSupport Map Generalization[J].Journalof Wuhan Technical U- niversityof Surveying and Mapping,2000,25 (1)35-41(, Delaunay [J]. [15]Elis W.ASourceBookofGestaltPsychology[M].,2000,25(1)35-41) [13]CortesC,Vapnik V.Support-vectorNetwork[J]. MachineLearning,1995, [14]ChangCC,LinCJ.LIBSVMA LibraryforSup- portvector Machines[J].ACM Transactionson InteligentSystemsand Technology,2011,2(3) LondonRoutledge & KeganPaul,1938 ArterialRoadsExtractioninUrbanRoadNetworks BasedonShapeAnalysis LUAN Xuechen 1,2,3 FAN Hongchao 4 YANG Bisheng 1,2 LI Qiuping 5 1 StateKeyLaboratoryofInformationEngineeringinSurveying,MappingandRemoteSensing, WuhanUniversity,Wuhan430079,China 2 EngineeringResearchCenterforSpatio-temporalDataSmartAcquisitionandApplication, WuhanUniversity,Wuhan430079,China 3 GuangdongRuituInformationSystemsCo,Ltd.,Foshan528305,China 4 ColegeofSurveyingandGeo-informatics,TongjiUniversity,Shanghai200092,China 5 SchoolofGeographyandPlanning,SunYat-senUniversity,Guangzhou510275,China AbstractThispaperpresentsashape-basedapproachtoextractarterialroadsfrom urbanroadnet- worksisappropriatetothehighlevel-of-detailfoundinvolunteeredgeographicinformation (VGI) datasets.firstly,the methodcalculatesseveralshapedescriptorsofeach polygonsurrounded by roads.secondly,somecandidatearterial-roadpolygonsareclassifiedasseedsfromalpolygonswith shapedescriptorsusingsupportvectormachine(svm).finaly,basedonseedpolygons,aregional growingmethodisproposedtoconnectthearterial-roadfeaturesaccordingtothegestaltconcept.an experimentshowsthatthismethodachievesgoodandefectivearterialroadextractionresults,reflec- tingtheconstructedskeletoninurbannetwork.furthermore,acomparisonanalysisofhigh-level roadsatributeswasalsocarriedouttoconfirmtheconsistencybetweenconstructedroadlevelsand ourdetectedarterialroads. KeywordsGIS-T;arterialroads;roadnetworks;Gestalt;roadhierarchies FirstauthorLUAN Xuechen,PhDcandidate,specializesindataanalysis,modelingandLoDrepresentationinurbanroadnetworks.E- mailxuechen.luan@whu.edu.cn CorrespondingauthorFAN Hongchao,PhD. hongchao@tongji.edu.cn FoundationsupportTheNational863ProgramofChina,Nos.2012AA12A211,2012AA12A204;theNationalNaturalScienceFounda- tionofchina,no ;thefundamentalresearchfundsforthecentraluniversities,nos , ;theguang- dongspecialfundforstrategicemergingindustrialdevelopment(high-endnew ElectronicInformation)Project,No ;the Scholarship AwardforExcelentDoctoralStudentGrantedby MinistryofEducation,No ;Wuhan UniversityPhD StudentShort-term AbroadResearchProjects.

NP [11] 7) 4)~ 6) ( ) : Kirk- patrick [26] 1 Fig.1 TextureMergingSketch Map ) 2) 3) 1 4) 5)

NP [11] 7) 4)~ 6) ( ) : Kirk- patrick [26] 1 Fig.1 TextureMergingSketch Map ) 2) 3) 1 4) 5) 40 3 2015 3 GeomaticsandInformationScienceofWuhanUniversity Vol.40No.3 March2015 DOI:10.13203/j.whugis20140231 :1671-8860(2015)03-0347-06 1 1 1 1 430079 : : :P208TP751 :A ISOMAP [1] [2] [25] [3] [4-10]