ongoingxyq.163.cbg.com

DYNAMICCARTOGRAPHICREPRESENTATIONOFSPATIO-TEMPORALDATAYangpinga,b*,TangXinmingab,WangShengxiaocaSchoolofResourceandEnvironmentScience,WuhanUniversity,129LuoyuRoad,Wuhan430079,China-yangping_whu@163.
com;bChineseAcademyofSurveyingandMapping,16BeitaipingRoad,Beijing100039-tang@casm.
ac.
cn;cInstituteofSpaceandEarthInfomationScience,TheChineseUniversityofHongKongRoom238B,SinoBuilding,Shatin,HK,CHINA-wsxiao@cuhk.
edu.
hkCommissionII,WGII/1KEYWORDS:Spatio-temporalDatabase,Spatio-temporalChangePattern,Dynamicsymbolbase,DynamicCartographicVisualizationABSTRACT:Aseffortsgrowtodevelopspatio-temporaldatabasesystemsandtemporalgeographicinformationsystemsthatarecapableofconveyinghowgeographicphenomenachange,itisimportanttoassociatethespatio-temporalprocesswithavividrepresentation.
Thepaperproposeawayforvisualizationbaseondynamicsymbolbase,bywhichrepresentthespatio-temporalprocessinamorevividandvisuallyway.
Theauthorsproposeanewclassificationofspatio-temporalchangepattern,whichbuildsafundamentalbasisfortherepresentationofdynamicphenomena.
Dynamicvisualvariableofdynamicsymbolforthedisplayofchangingspatial-temporalobjectiveswillbediscussed,aframeworkwillbedesignedandimplementedforthedynamicvisualization.
*YangPing,PhDCandidateinCartographyandGeographicInformationEngineeringofWuhanUni,Researchinterestsincluds:Spatio-temporalvisualization,GeophysicaldynamicvisualizationandTemporalGIS.
1.
INTRODUCTION1.
1DevelopmentinSpatio-temporalModellingOurplanetisdynamic.
Changesoccurconstantlyinallthecomponentsoftheearth'ssystem:itslithosphere,hydrosphere,biosphereandatmosphere.
Althoughhumaninterestintheworld'sdynamicsisnotnew,inthepastmajorfactorsthatinhibitedthestudyandunderstandingofthesechangeswerethelimitedavailabilityoflargetemporaldatasetsandalackofsuitablemethodsandtechniquestodiscoverpatterns,relationshipsandtrendsinsuchdata.
Spatio-temporaldatabasesaimtosupportextensionstoexistingmodelsofSpatialInformationSystems(SIS)toincludetimeinordertobetterdescribeourdynamicenvironment.
Inthepasttwodecades,alargeprojectcalledCHORONOROShadbeencarriedoutbyseveralagenciesinEuropeancountries,theobjectiveofistostudythedesign,implementation,andapplicationofspatio-temporaldatabasemanagementsystems(STDBMS).
CHOROCHRONOScoveringissuesrelatedtotheontology,structure,andrepresentationofspaceandtime;datamodelsandquerylanguagesforSTDBMS;graphicaluserinterfacesforspatio-temporalinformation;queryprocessingalgorithms,storagestructuresandindexingtechniques;andarchitecturesandimplementationtechniquesforSTDBMSswereresearchedunderthenetwork.
Asaresult,severalprototypeSTDBMSweredevelopedintheprojecttodemonstratetheinnovationsofmoretheoreticalwork.
Finally,applyingSTDBMSinrealisticproblemsettingsguidedtheresearchthroughouttheproject.
InChina,Adynamicfundamentalgeographicalinformationdatabaseisundertakenconstruction,alargemountof1:250000scalemapsheetshavebeentestedforthemanagementofmulti-versionspatialdata.
1.
2VisualizationofSpatio-temporalDataInordertopresentthecontentsofthedatacollectionortheresultsofqueries,mostmoderndatabasesystemsprovidetheuserwithvisualizationtools.
Someofthestandardtoolsforspatialdatabasesystemsincludebrowsers,plottersandmapdisplays.
Mostofthese,however,arepoorlyadaptedtodisplaydynamicand/ortemporalinformation.
Therefore,alternativegraphicalpresentationaltechniquesmustbeinvestigatedtosuccessfullycommunicategeographicalprocesses.
Toestablishguidelinesforspatio-temporalvisualization,wemustfirstdetermineourexpectationsandgoals.
Tothisextent,KraakandMacEachrenreviewaspectsofvisualizingdynamicprocesses.
Augmentinganexistingdefinition,theydefineatemporalmapas"arepresentationorabstractionofchangesingeographicreality:atool(thatisvisual,digitalortactile)forpresentinggeographicalinformationwhoselocationaland/orattributecomponentschangeovertime".
Spatio-temporaldatacanbegraphicallyrepresentedinmanyways.
Vasiliev(1997)proposedaframeworkforgraphicrepresentationinstaticmaps;Andrienkoetal.
(2003)attemptedtodothesameinanexploratorycontext,focusingoninteractiveanddynamicvisualizationofspatio-temporaldata.
FoundationforthevisualizationformalizationwaslaidbyBertinin1967(seeBertin,1974).
Hissemiologicalframeworkhasbeenfurtherelaboratedandextended.
Amongthemaindevelopmentsinthevisualdomain(seealsoSchlichtmann,1999)isresearchoneffectsofcombinationsofvariables(e.
g.
Spiess,1970).
Morrison(1974)andMacEachren(1994a,1995)TheInternationalArchivesofthePhotogrammetry,RemoteSensingandSpatialInformationSciences.
Vol.
XXXVII.
PartB2.
Beijing2008havearguedthatthevariablescolour(includingsaturation)andtexturearecompositesratherthanprimitives,andtheydefinedthedifferentcomponents.
Furthermore,newvariablessuchastransparencyandfuzzinessorcrispnessofsymboledgeshavebeendistinguishedandappliedtorepresentcharacteristicsofdataandofmetadata(MacEachren,1994a;vanderWeletal.
,1994).
AdaptationstoBertin'swayoflinkingthevariablestodatahavebeensuggestedaswell(e.
g.
Geels,1987;MacEachren,1995;Morrison,1974).
Kraak(1988)investigateddepthcues,i.
e.
variablesthatcanbeusedtosimulatethethirddimensionona2-Dsurface,whileDiBiaseet.
al.
(1992)andMacEachren(1994b)proposeddynamicvariablesforanimatedrepresentations,togetherwithsuggestionsfortheiruse.
Characteristicsofgeographicdatacanalsoberepresentedbysignsorsignalsformodesofperceptionotherthansight.
Vasconcellos(1993)madeanattempttodefinetactileequivalentsofBertin'sgraphicvariablesforvisuallydisabledpeople.
Morerecently,Griffin(1999)addedkinaestheticvariables(e.
g.
resistance,friction)forapplicationinfullyimmersiveVirtualEnvironments.
BarendKbbendescribedpreliminarytestsforevaluatingtheperceptualpropertiesofdynamicvisualvariables.
Bythecomparisonofsixkinddynamicvisualvariables,includedmoment,duration,frequency,order,rateofchangeandsynchronization,theygaveoutthatdynamicvisualvariableswouldonlyrenderfavorableresultsintheuseofcartographicanimations.
Thesedevelopmentsshowthattheframeworkfortherepresentationofgeodatabyperceptualvariables,originallyfoundedbyBertin,isadynamicconstructthatisstillevolving(Blok,1998,Fairbairnetal.
,2001;MacEachren,2001).
Importantreasonsforongoingdevelopmentsareincreasinginsightandprogressintechnologicaladvancements.
Itisclearthattherepresentationframeworkisnotafixedconstruct;itwillbefurtherextendedandadapted.
Blokdevelopedtheframeworkbyintegratingtheformervisualvariablesintofourcategories:momentofdisplay,order,duration,frequency,gaveouttherelationshipsbetweenthedynamicvisualizationvariables,providedinteractionsoptionsfromadesignperspective,madetheutilizationofthevariablesfromauserperspective,andtestedtheeffectsoftheusinginteractionsofvariousdynamicvisualizationvariables(blok,2005).
Suchasweknow,AmountsofworkhavebeendoneontheevolvingoftheBertin'sframework,butmoreshouldbedoneforfurtherresearch.
Thegoalofthispaperistoproposeanewframeworkwhichcanvisualizethespatio-temporalprocessesrefertoeachclassificationoftemporalphenomena.
Bytheutilizingofdynamicsymbolbase,amethodologyfordynamicvisualizationofchanginggeographicdatawillbecarriedout.
2.
SPATIO-TEMPORALCHANGEPATTERN2.
1WhatisImplyinSpatio-temporalDataAsweknow,theworldwearelivinginisaneverchangingplanet,informationaboutchangingarecontainedinthedatasetthatcollectedindifferentversion.
Spatio-temporaldatabasedestinedtodescribeandmanagehistoricalgeographicalphenomena.
However,describingreal-worldevolutionisacomplextask.
Onemayobservethestatusofentitiesbeforeandafterachangeoccurs,thesearefactsandconsequences.
Aneventisasetofrelatedchangesleadingtoanewstatus.
Eventsmaybeobservedwithoutknowledgeaboutthemechanismsleadingtochange.
However,wepostulatethatchangehappenwhenasetofactiveentitiesorforcestransformtheirenvironment.
Mostofthetimethesetransformationsdonotoccuratrandombecausetheyareconstrainedtopreviousstatusandobeytoevolutionlaws(causalrelationships).
Discoveringtheselawsistheultimategoalofscience(explanation).
AknowledgeofSpatio-temporalProcessinnecessaryforvisualizationofthedynamicdata2.
2Spatio-temporalProcessVariousmodelsofchangehavebeendevelopedbymathematicians,geographers,philosophers,andcomputerscientists.
KathleenHornsbypresentedanapproachtospatio-temporalknowledgerepresentationbasedontheexplicitdescriptionofpossiblechangestogeographicphenomenamodeledatahighlevelofabstractionasidentifiableobjects.
Startingwithasetofbasictypesofchangewithrespecttotheexistenceandnon-existenceofobjectswithidentity,amethodologywaspresentedthatsystematicallybuiltonthesefundamentalconceptsandderivedfurthertypesofchangethatwerepossible.
Thischange-basedmodelprovidedabetterunderstandingofthesetofpossiblealterationstowhichanobjectcanbesubjectasitevolvedoverspaceandtimeandenabledtheextensionofspatialdatamodelsandthedevelopmentofGISquerylanguagesthatincorporatesuchsemanticsofchange.
Althoughnoexplicitlyspatialinformationhasbeenincorporatedinthismodelofchange,ithasbeenshownthattrackingchangestoanobject'sidentityoverperiodsofexistenceandnonexistence,givesusefulinsightsintothebehaviorofanobjectovertimethatarerelevanttomanycasesofspatio-temporalchange.
ChristopheClaramuntproposedastandardwayofdesigningSTPs,aimedvalidforeverydatamodelthatsupportsthetimeandspacedimensions.
HepresentedataxonomyofbasicSTPs,extendedthespatio-temporalprocessedintwoways,andderivedacompleteclassification,includingendogenousapproachandexogenousprocesses,consistedofTranslation,SuccessionandPermutation,Re-allocation,SplitandUnion,Production,ReproductionandTransmission.
introducedrelateddatabasemodelingissues,expressedthreedesignpatternsinaspatio-temporalobject-relationshipdatamode(ChristopheClaramunt1997).
2.
3QualitativeRepresentationoftheevolvingspatialentitiesQualitativereasoning,whichiswidelyusedtodefinespatialrelationships,mustthereforebeextendedtothetemporaldimensiontoprovideasetofwell-defined,homogeneousspatio-temporalrelationships.
Theevolutionofanindependentspatialentityinvolveschangesinitssizeandshapeaswellasmovementprocesses.
Themulti-linearevolutionrepresentschangesofasetofspatialentitiesthatareofthesamespatialtypeandthatcoversthesamearea,theidentifiedprocessesarethesplitofaspatialentityandtheunificationorre-allocationofseveralspatialentities.
ThepropertiesoftopologicalspaceshavebeenextensivelyusedasamathematicalsupportforGISdesign.
Theyprovideaformalsupportfortheexpressionofspatialrelationships(Alexandroff1961).
Atemporalmetricassignsvaluestothingsintimeandmeasuresdurationwhereasatemporaltopologyallowsthestudyofpropertieswhicharepreservedunderallcontinuoustransformations.
TheInternationalArchivesofthePhotogrammetry,RemoteSensingandSpatialInformationSciences.
Vol.
XXXVII.
PartB2.
Beijing2008Ajudgeonchangeofobjectinthefour-factormodel:theobject,objectproperties,objectorattributethestatusofthepropertyvaluelogicalresult:Change(O)=F(O,P,S,R,Tn).
(n=0toN)WhichOmeansObject,Pmeansobjects'properties,Smeansthestateattributes,Rpropertyvaluemeansthelogicalcomputationresult,Tnmeansthetemporalinformationwhichisoftenrepresentedinspottime.
Inotherwords,thegeographicalchangesofobjectsareconcernedwithobject,objectproperties,timeandstateproperty.
Aalgorithmcanbedevelopedforthedetectionofchangebytheaboveregulations.
Indifferentspottime,maybesomeoftheaspectsaredifferentfromtheprevioustime,thenitcanberecordedbydatabase,andeachkindofchangecanbeassociatethedefinedspatio-temporalchangepattern.
Inthefollowingsection,spatio-temporalchangepatternsaretobeproposedfortheimplementationofdynamicrepresentationofspatio-temporaldata.
2.
4Spatio-temporalChangePatternInthispaper,weproposeapracticalclassificationofSpatio-temporalchangepattern,abbreviatedasSTCP.
Forthereasonofactuallyimplementation,wegetusedtotheconceptionofLayers,Suchasweknow,Layersarewidelyusedfortheconstructionofgeographicaldata,whenweconstructthespatio-temporaldatastorageinthedatabase,therearethreefeaturespatterns,includespoints,linesandpolygonswithinalayer,foreachgeographicalobjectorphenomenon,itisabletorepresentinthethreepatternsoffeatures.
Onthebasisofthreepatterns,classificationcorrespondingeachpatternisproposedtodescribetheprocessofchangeofspatio-temporalobjects.
Hencewedrawupaspatio-temporalprocesstaxonomywhichisusedforthetaxonomyofSTCPbytheanalyzingofessenceofphenomenachange,itcanbeshowedinthefollowinggraph:Figure1.
AgraphfortaxonomyofSTCPTheprocedurethatfromstarttoendhasnodifferenceexceptthetimevariablehasbeenchanged,wenameitaNoChangeprocess,thisprocessiseasilyignored,butintherealdatabase,therecordaboutthisinformationissolidlystoredintheSpatial-temporaldatabase.
ApartfromtheprocessofNoChangeweproposeadetaildefinitionofthespatio-temporalchangepatterns,transformationoccurredduringtheSpatial-temporalprocesscouldbereflectedinmanyways,inthepaper,weplantoclassifiedthemintwoaspects,oneofthemoccurredinthegeometricaspect,wegiveitanameGeometricChange;Anothercanbehappenedinthematicway,whichcanbenamedafterThematicChange,thereasonforthisclassificationliesinalogicaldesignpatternofdatabase,thereisacommonwayforthedatastoringstrategythatthegeometricdataandpropertydataareseparatelydepositedbutwithaweakrelationbytheusingofkey.
RefertotheGeometricChange,weproposeafurthermoreclassificationwhichcomprisingpatternBorn,Grow,Shrink,NormalChangeandDie,Inordertodealwithchangeconcerningtwoormoreobjects,weproposedExpandandMergefortheGrowpattern,andContractandSplitfortheShrinkpattern.
ForBorn,foreachspatio-temporalobject,thecreationisthemostandtheverybeginningstepduringit'slifecircle,itisanoppositeprocessofDIE,meansthatanewlyobjectsexistafteradeathofothers,alsohasmanyinstances.
ForGrow,thereexiststwosituations,acommonphenomenonthataspatio-temporalobjectenlargeit'ssizeduetoalotofreasons,wenameExpandtorepresentthiskindofprocess.
Anotherissueisoccurredamongtwoormoreobjects,agrowingspaceoccupiedbyanobject,thisobjectiscomprisedbytwoormoreobjects,thesespatio-temporalchangepatternsnamedafterMergewhichcanbederivedbytheIDnumberitcorrespondto.
ForShrink,WeproposetheContracttodescribewhenaobjectdecreasingit'sspacethatoccupiedinthegeo-spatialway,asaresultofContractprocess,geometricalareaorvolumereducedinacomputableway,Contractisaoriginalprocess.
Asfortwoormore,WeproposeSPLITasabasicchangepatternforthedescriptionofchangeoccurredinvolvingseveralspatio-temporalobjects,fortheSplitprocess,atthebeginning,thereisoneobjecthasanintegralpartofgeometricproperties,intheend,newlybornobjectshavebeencreated,asaresult,thespacetheoccupiedbytheoriginalobjectshasbeenreducedthroughthisprocess.
Normalchangeisaspecialchangepatternwhichisdistinctfromtheaboveprocess,thesizeorspacehasnotbeenchangedbutothercharacteristicfeaturessuchasshape,locationandobjectIdbecomedifferent.
Itisacomprehensivechangethatcanbereflectedinmanygeometricalaspects,thereisaexample,fortheSplitchange,ifthespacehasnotchangedduringtheprocess,itturnedintoNormalchangeinsteadofShrink.
Ifaobjectchangeitlocationasawholebody,normalchangehappenedinthiscase.
TheLastprocesswenameitDie,theprocessisregardedasaterminationofaspatio-temporalobjects,TheDieprocesscanbedescribedinsuchaway,theidentityoftheobjecthaschangeintoanotherorthegeometryhasvanished,theprocessisacreationofthenewobjects,leadingtothecontraryprocessofBorn.
TheclassificationofSTPchangepatterndobenefitfortherepresentationofchangingdata,andbuiltafundamentalbasisforthevisualizationofSTP,inthefollowingparagraph,numerousvariableswillbeutilizedforthedisplayingoftheSTP,withrespecttoeachprocessundertheframeworkproposedhere.
TheInternationalArchivesofthePhotogrammetry,RemoteSensingandSpatialInformationSciences.
Vol.
XXXVII.
PartB2.
Beijing20083.
DYNAMICVARIABLESANDTEMPORALMAPPINGIthasbecomeclearthatthetraditionalvisualvariables,whichwewillcallthestaticvisualvariablesfromhereon,donotsufficeindescribingtheaddedmeansofexpressionwehaveindynamicvisualizationofspatio-temporalprocessedthroughcartographicanimations.
Tothisendsix"new"visualvariablesincludingmoment,duration,frequency,order,rateofchange,synchronizationhavebeenintroducedbyDiBiaseetal.
(1992)andMacEachren(1994).
Thesewillbecalledthedynamicvisualvariablesfromhereon.
ResearchbyDiBiaseetal.
(1992)andKoussoulakou&Kraak(1992)hasshownthatvisualvariablescanindeedbeusedontheindividualframesofananimationinsuchawaythattheseimageseffectivelycommunicatethecartographicmessagetotheuser,whilethemovementoftheanimationgivesthemessageanextradimensionand"newenergy".
Furthermore,thefindingsofKoussoulakou&Kraak(1992)showedthatusinganimatedmapshelpedusersgraspthecontentsofamessageinamoreeffectivemannercomparedtousingtraditionalstaticmapsormapseries.
3.
1FourImplementalDynamicVariablesInthisparagraph,inordertogetanimplementalframeworkofdynamicvisualizationfortheSTP,wechoseMoment,Duration,OrderandFrequencyfortherepresentationoftemporalaspectofobjects,therearethereasonsforthechoice:3.
1.
1MomentInmosttime,welookprocesstimeasacontinuousvariable,however,eachprocesscanbedividedintodiscretetimepieces,wecallitMOMENT,ThemomentthatanelementinthemapchangesduringaSTPcanbeusedfortemporalaswellasnon-temporalanimations.
e.
g.
everyprocedurehasabegintimeandendtime,somekeyremarkableMOMENTareoftenrecordinspatio-temporaldatabase.
3.
1.
2DurationFromtheStarttotheend,thetimespanthatbetweentwodifferentMOMENTisnamedDURATION,wecanalsonameitafterfromonespecificstatetoanotherstateofobjectsorphenomena,Thelasttimealwaysisanimportantiteminthegeographicinformation,e.
g.
howlongwillbetyphoonlast.
3.
1.
3OrderOrderisusedtodifferentiatechangesequenceamongseveralspatio-temporalchanges,sometimesweuseaframeaccordingtoaspecifiedspatio-temporalprocess,Theorderofchangeissimilartothetopologicalrelationshiporspatialneighborhood,inonehand,thesequenceoftheframecanberepresentedbyORDER,intheotherhand,orderalsoreferstotheorderofphasesinaseriesofchangesinthespatialdomain.
3.
1.
4FrequencyFrequencyreferstothenumberoftimesthataphaseisrepeatedinaseriesofchangesinthespatialdomain.
ItisalsousedtodeterminehowtheframeforrepresentingtheSTPtoberepeated,sincealotofphenomenareoccursaccordingtoaregulartime,herethefrequencycanconceptualizethisphenomenaverywell.
Altogether,thevariablementionedabovemaynotsufficienttodescribealltheSTPrepresentingtheobjectsandphenomena,howeveritcanbuildaimplementalframeworkforapracticalimplementation.
3.
2TemporalMappingThereareatleastthreekindoftimeforthespatio-temporalobject,RealWorldTime,DatabaseTimeandDisplayTime.
whendealingwithatemporalprocessadirectrelationbetweendisplaytimeandworldtimeexists.
WorldTimeisthetimescaleofreality,themomentaneventtakesplaceintherealworld.
ExamplesoftheseprocessesarethoseoftheDutchcoastlinefromRomantimesuntiltoday,boundarychangesinAfricasincetheSecondWorldWar,orthechangesofyesterday'sweather.
Timeunitscanbeseconds,weeks,oryears.
Inordertorepresentthespatio-temporalprocessinadynamic(animated)way,themappingfromRealWorldTimetoDisplayTimeisnecessary,becauseitisimpossibletodisplayatimesequencechangeaccordingtoactualtime.
Figure2showarelationbetweenRealWorldTimetoDisplayTime,T1andT2areworldtime,DisplaytimeT1'andT2'arederivedthroughamapping.
AllthistimearerecordedintheDatabasetime,thatmeansthatDatabaseTimeisacompleterecordtemporalvariableaggregation.
NotonlythetransactionaltimebutalsotheRealWorldTimeandDisplayTimearederivedfromit.
Figure2.
MappingfromRealWorldTimetoDisplayTime3.
3CombinationofTraditionalandDynamicVariablesBarendKbbenandMustafaYamanhavecarriedoutapreliminarytestforevaluatingtheperceptualpropertiesofaseriesofdynamicvisualvariables,Thetestresultsshowsthatthedynamicvisualvariablemomentappearstobenotveryusefulforcartographicanimationsunlessincombinationwiththestaticvisualvariablecolourandsomeothervariables.
Perhapsthemostimportantconclusionhastobethatselection,whichissoimportantinmakingcartographyastrongtoolforcommunication,cannotbeeffectivelyattainedbyusingdynamicvisualvariablesalone.
Thetraditionalvariablesinordertoaccompanydynamicvariablesareasfollows:TheSIZEofanareaonamapmaybechangedtoshowchangesinvalue,itrepresentsthesizemodificationsoftheobjectsduringaspecifiedtimeinterval.
Forexample,thesizesofcountriesaremadeproportionallylargerorsmallertodepicttheamountofoilorcoalreserves.
Ananimationcanbeusedtotransformthemapofoilreservesintothemapofcoalreservestoshowthedifferencesinlocationofthereserves.
SHAPErepresentsauniquecharacterofeachobject,sometimetheshapecantellmuchofthequalityofaobject.
Anareaonamapcanbemadetochangeinshape.
Theshape(andsize)ofGreenlandvariesasaresultoftheinfluenceofamapprojection.
AnSTPcanbeusedtoblendbetweenthetwoshapestoaccentuatetheeffectofthedifferentprojections.
COLOURisusedtoconceptthecalorificpropertiesthatobjectstoberepresented.
Acolortransitionofsptio-temporalobjectsmeansfromonecolourtoanotherduringaspecifiedtimeintervalforacertainnumberofiterations.
Thecolouranimationcanbeusedtohighlightobjectsofinterestonamap.
Theblinkingofpointsymbolgiveasenseofstressing.
OPACITYrepresentstheanimationvisibility,inotherwords,thetransparencyoftheobjectsfromonedegreeoftransparencytoanotherduringaspecifiedtimeinterval.
ThetypicalusagesTheInternationalArchivesofthePhotogrammetry,RemoteSensingandSpatialInformationSciences.
Vol.
XXXVII.
PartB2.
Beijing2008oftheopacityanimationareblinkingtodrawattention,fade-intomakeobjectsprogressivelyappear,andfade-outtomakeobjectsprogressivelydisappear.
DIRECTIONrepresentsthedirectionchangeoftheobjectsfromastartingangletoanendingangleduringaspecifiedtimeinterval.
Objectsmaybemodifiedwithdirectionchangeintermsofapositionpoint.
Thedirectionanimationisusedtorepresentthesymbolswiththeobviousarrowordirection.
AnothervariableDENSITYisoftenusedtorepresentthedegreeofscatterorassembleofspatio-temporalobjects,densitybecamelargeriftheobjectsbecamemorecollective,whereasitbecamesmallerwhentheobjectscattersparsely.
Byassigningthetraditionalvariableswithdynamicvariables,thecombinationofthembecomeflexibletovisualizechangingphenomenonandalteringobjects.
Thenaframeworkisneededtobecreatedforthiscombination,whichbasedonthedynamicsymbolbase.
4.
DYNAMICVISUALIZAITONBASEDONDYNAMICSYMBOLBASEAbovewehavementionedmanyaspectsofdynamicsymbol,whichisusedtovisualizethespatio-temporalprocess,inthissection,thedynamicsymbolbaseisproposedtoestablishanstrategyfortheimplementationofdynamicsymbol.
Thestrategyareproposedtodefineaprocedurewhichcanmakeamulti-versioneddataintoadynamicvisualization.
First,theSpatio-temporalassociationwillbeusedtobuildahistoricalrelationshipbetweendifferentspatio-temporaldata,thendifferentSpatio-temporalprocessesaregotandthemappingoftimearederived,basedonthedynamicsymbolbase,temporalsymbolandnon-temporalarecreatedbythecombinationofdynamicvariablesanddynamicvariables.
Thusdynamicvisualizationcanbedisplayedbyacontroloverthetime-seriesdynamicmapthroughavisualinterface.
4.
1Spatio-temporalAssociationOnthebasisoftheclassificationofspatio-temporalprocesses,adoptingaseriesofalgorithmsfordetectingchangesbetweendifferentdataversionsthroughgeometryandthematiccharacteristicsofspatio-temporalobjects,differentkindsofSTPwillbederivedfromtheassociation.
Atimemappingalgorithmwillbeusedtogettherelationshipbetweenrealworldtimeanddisplaytimeaswell.
4.
2UtilizationofDynamicSymbolBaseHereweproposeadynamicsymbolbasefortheimplementationofdynamicvisualization,whichprovidesamethodologyratherthanaconcreteentityfortheapplication.
Apartfromofferingabaseforthetraditionalmechanismforstaticgeographicaldata,thedynamicsymbolbasebuiltupaframeworkfortwokindsofdynamicsymbolasFigure3:4.
2.
1TemporalSymbolLotsofspatio-temporalprocessesaresuccessiveproceduresthatcanbedisplayedinacontinuousrepresentation,inthepaper,weadopttemporalsymboltovisualizesuchkindofsuccessiveprocesses.
Bythecombinationofdynamicvariablesandtraditionalvariablessuchascolour,size,opacityandsoon,theprocesscanbedisplayedwell.
Forexample,urbanizationcanbeshowedvividlybytemporalgraduallysize-changedpolygon,etc.
Anexampleofthetemporalsymbolistheadditionofsubsequentmaplayers,whichleadstheviewerthroughatheme,tohelpunderstandspatialandcontextualcoherence.
Thisleavesthistypesofcartographicanimations,whicharethecoreoftheclassificationpresentedhere.
Itistherelationbetweenspatialdata'scomponentsanddisplaytimewhichdistinguishesthemfromeachother.
Figure3AFrameworkforDynamicVisualizationBaseonDynamicSymbolBase4.
2.
2Non-temporalSymbolAnumberofphenomenonarenotsostrictlyrelatedtothewholeprocessofapatio-temporalobjects,however,somecharacteratspecificmomentneedtobehighlighted.
Non-temporalsymbolisnotstrictlylinkedwithworldtime,thedisplaytimeofitisappliedtoexplainspatialrelationsbypresentingindividualimagesinlogicalsequence.
Sometechniqueswillbeusedtodemonstratehighlightsoremphasisofspatio-temporalprocesses,suchastheblinkingandflashingofpointsymbol.
Anotherexampleofthisistoshowadatasetindifferentgraphicrepresentations,suchasanisolinemap,ansmoothstatisticalsurfaceoradotmaptoprovidethevieweracomprehensiveimpressionofthesamedataset.
4.
3VisualInterfaceforDynamicControlFortheuserofadynamicsymbol,itisimportanttohavetoolsavailablethatallowforinteractionwhileviewingtheanimation.
Seeingtheanimationplaywilloftenleaveuserswithmanyquestionsofwhattheyhaveseen.
Justare-playisnotsufficienttoanswersquestionslike'Whatwasthepositionofthecoastlineinthenorthduringthe15thcentury'Mostgeneralsoftwaretoviewanimationsalreadyofferfacilitiessuchas'pause',tolookataparticularframe,and'(fast-)forward'and'(fast-)backward',togotoaparticularframe.
Moreoptionshavetobeadded,suchasapossibilitytodirectlygotoacertainframebasedonforinstanceatemporalquery,ortheabilitytore-orderingindividualframesbasedonaattributequery.
Thisbecomesespeciallyrelevantifwerealizethatanimationwillnotonlybeusedtopresentspatialdata,butwillbeincreasinglyusedinanexploratoryenvironment.
Insuchanenvironmenttheanimationisjustoneofthealternativeviewonehasavailabletostudythedataathand(Dykesetal.
,2005).
Wedesignedavisualinterfacethathelpuserstomanipulatetheprocessofdisplaying,throughwhichpeoplecanplaybackandquerysomedetailsoftheSpatio-temporalprocess.
Aspeedcontrolfunctionoftheprocessalsoprovidedfordifferentvelocityanalysis.
UsercangetavividimpressionofthechangeTheInternationalArchivesofthePhotogrammetry,RemoteSensingandSpatialInformationSciences.
Vol.
XXXVII.
PartB2.
Beijing2008Masuoka,P.
,Acevedo,W.
,Fifer,S.
,Foresman,T.
Tuttle,M.
,1996.
TechniquesforvisualizingurbangrowthusingatemporalGISdatabase.
PresentationattheASPRS/ACSMAnnualConventionandExhibtion,Baltimore,MD,pp.
22-25.
thathadhappenedfromthedatawhicharestoredinthespatio-temporaldatabase.
4.
4TestandAnalysisAtestoftheimplementationoftheframeworkiscarriedoutbasedontheresidentialexpansionofacityofBAOJIinthenorthwestofshananxiprovinceinChina.
ThefollowinggraphdiagramshowsEntity-basedspatial-temporalprocessdisplayingbasedondynamicsymbolbase.
Anurbangrowingprocesscanbeshowedbyaserieschangethroughdynamicsymbol.
ZHUGuorui,2004.
Cartography.
WuhanUniversityPress,pp.
34-35.
AITinghua,1998.
Dynamicsymbolanddynamicmap.
JournalofWuhanTechnicalkUniversityofSurveyingandMapping.
23(1),pp.
47-51TANGXinming,WUlan,1999.
Spatio-TemporaldatamodelandGeographicInformationSystemFrame.
RemoteSensingInformation01,pp.
4-8.
WULixin,SHIWenZhong.
,2003.
PrinciplesandAlgorithmsofGIS,SciencePress,Beijing,pp.
154-156.
HuangXingyuan,HUPeng,2001.
TutorialofGIS.
WuhanUniversityPress,pp.
204-206.
T1T2Figure4TimeserieschangeshowedbytemporalsymbolVasiliev,I.
R.
,1997.
MappingTime.
Monograph49,Cartographica34.
2.
5.
CONCLUSIONDynamicSymbolofferstheusertheopportunitytoseeandquerychangesspatialpatterns.
Dependingonthenatureofthedataonecanapplydifferentdesigntechniques,orchangetheviewpointonthedata.
DynamicSymbolwillofferabetterinsighttomappedphenomena.
However,thiswillonlyworkwhentheuserenvironmenthastheproperoptionsforinteraction.
Inanexploratoryenvironmentthedynamicsymbolwillbeoneofthestrongalterativeviewsonthedatathatsupportsknowledgediscovery.
Andrienko,N.
,Andrienko,G.
,Gatalsky,P.
,2003,Exploratoryspatio-temporalvisualization:ananalyticalview.
JournalofVisualLanguages&Computing14,pp.
503-541.
MacEachren,A.
M.
,1994a.
Sometruthwithmaps:aprimeronsymbolization&design.
WashingtonD.
C.
:AssociationofAmericanGeographers.
MacEachren,A.
M.
,1994b.
TimeasaCartographicvariable.
In:H.
Hearnshaw&D.
Unwin,eds.
,VisualizationinGIS.
London:Wiley&Sons,pp.
115-130.
ACKNOWLEDGEMENTTheauthorswishtothankChinaNationalHigh-techR&DProgram(ProjectNo.
:2006AA12Z214)andChinaNationalFundamentalSurveying&MappingProject:(ProjectNo.
:1460130524207)andChineseAcademyofSurveying&MappingBasicScientificResearchProject:(ProjectNo.
:G7724)andDigitalMapping&LandandResourceApplicationEngineeringStateBureauofSurveyingandMappingOpenFundProject(ProjectNo.
:GCWD200703)forfinancialsupport.
DiBiase,D.
,MacEachren,A.
M.
,Krygier,J.
B.
&Reeves,C.
,1992.
Animationandtheroleofmapdesigninscientificvisualization.
CartographyandGeographicInformationSystems,19(1),pp.
201-214,265-266.
Vasconcellos,R.
,1993.
Representingthegeographicalspaceforvisuallyhandicappedstudents:acasestudyonmapuse.
In:P.
Mesenburg,ed.
,Proceedingsofthe16thInternationalCartographicConference,Cologne,German,InternationalCartographicAssociation,2(1),pp.
993-1002.
Griffin,A.
2003,Anempiricalstudyofinteractivemapsinthecontextofageographicalsimulationmodel.
In:Cartographicrenaissance.
Proceedings21stInternationalCartographicConference,Durban,SouthAfrica.
InternationalCartographicAssociation,pp358-364.
REFERENCESKoussoulakou,A.
andKraak,M.
J.
,1992.
Spatio-TemporalMapsandCartographicCommunication,TheCartographicJournal,29(1),pp.
101-108.
Dykes,J.
,1996.
Dynamicmapsforspatialscience:aunifiedapproachtocartographicvisualization.
In:D.
Parker,ed.
,InnovationsinGIS3.
Bristol,Pennsylvania:Taylor&Francis,pp.
177-187.
Abraham,T.
andRoddick,J.
F.
,1999,Surveyofspatio-temporaldatabases,GeoInformatica3(1),pp.
61-99.
Kbben,B.
,Yaman,M.
,1996.
EvaluatingDynamicVisualVariables.
In:Ormeling,F.
J.
,Kbben,B.
,PerezGomez,R.
(eds.
):ProceedingsoftheSeminaronTeachingAnimatedCartography.
InternationalCartographicAssociation,Utrechtpp.
45–51.
Kraak,M.
J.
andKlomp,A.
,1995.
"AClassificationofCartographicAnimations:TowardsaToolfortheDesignofDynamicMapsinaGISEnvironment".
TheSeminaronTeachingAnimatedCartography,Spain,http://cartography.
geog.
uu.
nl/ica/Madrid/kraak.
html.
El-Geresy,B.
A.
,Abdelmoty,A.
I.
andJones,C.
B.
,2000.
EpisodesinSpace:QualitativeRepresentationandReasoningoverSpatio-TemporalObjects.
InternationalJournalonArtificialIntelligenceTools,IJAIT,WorldScientificPublishing,9(1),pp.
131-152.
KoussoulakouA.
andKraakM.
J.
,1992.
Spatio-TemporalMapsandCartographicCommunication,TheCartographicJournal,29(1),pp.
101-108.