DEM 1(d), DEM, 6km - 803m, 611m DEM GIS, 2 m 3, 1 Fig.1 ExperimentData ;W ;k 2 X,k (3) C 0 C 1 C 2, [5] (simulatedannealing,sa), X, : C 0 C

Size: px

Start display at page:

Download "DEM 1(d), DEM, 6km - 803m, 611m DEM GIS, 2 m 3, 1 Fig.1 ExperimentData ;W ;k 2 X,k (3) C 0 C 1 C 2, [5] (simulatedannealing,sa), X, : C 0 C"

笃历鞠
5 years ago
Views:

1 GeomaticsandInformationScienceofWuhanUniversity Vol.40No.3 March2015 DOI: /j.whugis : (2015) , ,, ,, : DEM GIS 690 m,,, : ; ; ; :P231.5;P208 :A, 4 20 GIS, 5 12 DEM GIS,,,, GIS,,,, 1 [1-4], 5 12, GIS 1 5 DEM ( 1) 3S,,DEM, : : ( , ); (CKSF /KJ); ( ) :,, sunbathing@163.com.

334 2015 3 DEM 1(d), DEM, 6km - 803m, 611m DEM GIS, 2 m 3, 1 Fig.

2 DEM 1(d), DEM, 6km - 803m, 611m DEM GIS, 2 m 3, 1 Fig.1 ExperimentData ;W ;k 2 X,k (3) C 0 C 1 C 2, [5] (simulatedannealing,sa), X, : C 0 C 1 C 2, I 1 +I ( 2 O1 +O2 Δt- 2 ) ( 2 ) Δt= ΔW (1) (0,1) C 0 C 1 C C 0 -C 1 W =k[xi+ (1-x)O]=kQ (2) (3) (1) (2) : O 2 =C 0I 2 +C 1I 1 +C 2O 1 (3) O 2, J 烄 C 0 = 0.5Δt-kx (Q(i)) 0.5Δt+k-kx C(i) C(i+1), 烅 C 1 = 0.5Δt+kx J(Q(i+1)) de = J(C(i+,C 0 +C 1 +C 2 =1 0.5Δt+k-kx 1))- J(C(i)) 0, C 2 = -0.5Δt+k-kx C(i)= C(i+1), C(i+1) 烆 0.5Δt+k- kx i+1 (4),I(input) O(output)

3 40 3 : 335,, exp(-de/t) , ( 1) C 0 C 1 C 2 2, ~ ,, 2 Fig.2 Muskingum ParametersSolutionBasedonSimulatedAnnealing 1 Tab.1 EquationParametersandAccuracyValidation C0 C1 C2 /(m 3 s -1 ) /(m 3 s -1 ) /% Jmin= Jmin= ,

4 Fig.3 RunofversusElevationCurve h 1 h 2, h 0 [6-9] GIS DEM c Δt Δx <1 (6),c ;Δt ;Δx,,, DEM, m Saint-Venantequations, Q h :, B h ( ) 烄 t + Q x =ql ( ) 烅 Q t + x ( α Q2 A ) +ga h [10-13] x =ga (S 0 -S f ) 烆, (5), 690 m,q A h ;, t x ( );g α 4 ( α 1.0);S 0 S f ;ql,,, 4.1, [15] TVD-, MacCormack,, -, TVD, 3.2 m 3 /s, 4 m 3 /s, 2 m 3 /s CFL, Courant,, :, 4

5 40 3 : m Fig.4 ProcessofRunof,ElevationinEntirelyColapseModewhenLevelis690m Q 0=82000m 3 /s,, 76 m 3 /s, O 1 O 2, I 1 I 2 2 ( 1h ), 2 Tab.2 CalculationResultsBetweenTangjashanandXiangshuiStation /h I2/(m 3 s -1 ) I1/(m 3 s -1 ) O1/I 1/(m 3 s -1 ) O2/I 2/(m 3 s -1 ) O 1/(m 3 s -1 ) O 2/(m 3 s -1 ) , X Y ( m 3 /s m 3 /s), m m 3 690m m m,, GIS, DEM H=690m, Q 0 =82000 m 3 /s,, 5 ArcGIS DEM, DEM, 4.2, 690 m shpfile, ArcGIS,,,, 6

6 m Fig.5 InundatedAreaAcquiredby Muskingum MethodwhenDam BreachElevationis690m m Fig.6 InundatedAreaAcquiredbyHydrodynamicsMethodwhenDam BreachElevationis690m 7 DEM,, ;, DEM DEM,, DEM,DEM DEM,, (530 m ) DEM, DEM,, 590 m DEM 60 m, DEM,,

40 3 : 339 7 Fig.7 ComparativeInundatedAreaBasedonTwoPrediction Methods Dam[J].Natural Hazards,2013,65:41-62 [7] Ehrlicha,D.etal.IdentifyingDamageCausedby [1] Pan Shibing,Li Xiaotao.

7 40 3 : Fig.7 ComparativeInundatedAreaBasedonTwoPrediction Methods Dam[J].Natural Hazards,2013,65:41-62 [7] Ehrlicha,D.etal.IdentifyingDamageCausedby [1] Pan Shibing,Li Xiaotao.Using Remote Sensing the2008 WenchuanEarthquakefrom VHR Remote Datato Monitor Dammed Lakes Caused by Wen- SensingData[J].InternationalJournalof Digital chuanearthquake[j].journalof Geo-Information Earth,2009(2): Science,2009,11(3): (, [8]. ZhaiYifeng.TheFloodDisasterAssessmentModel 5 12 [J]. BasedonGIS/RS[J].Yelow River,2003,25(3):,2009,11 (3): ) 6-7(. GIS/RS [2] [J].,2003,25(3):6-7) FanJianrong,TianBingwei.InvestigationonDam- [9] ming ObjectInduced bythe Earthquakeof Wen- TolgaG,FanXuanmei,Van WestenC.etal.Dis- chuanon May12Basedon Multi-platform Remote tributionpaternofearthquake-induced Landslides Sensing[J].Journalof Mountain Science,2008, Triggeredbythe12 May2008 Wenchuan Earth- 26(3): (,. quake[j].geomorphology,2011,133: [J]. [10]Rahman M,Chaudhry M H.SimulationofDam-,2008,26(3): ) breakflow withgrid Adaptation[J].Advancesin [3] CuiPeng,Zhu Yingyan, Han Yongshun,etal. Water Resources,1998,21 (1):1-9 The12 May Wenchuan Earthquake-induced Land- [11]XueYang,Xu Weilin,etal.ExperimentalStudyof slidelakes:distributionandpreliminaryriskeval- Dam-breakFlowinCascadeReservoirs withsteep uation[j].landslides,2009(6): Botom Slope[J].Journalof Hydrodynamics, [4] HsuY S,Hsu Y H.ImpactofEarthquake-induced 2011,23(4): DammedLakesonChannelEvolutionandBed Mobili- [12]Long Wenfei,ZhangXinhua.StudyonthePredic- ty:casestudyofthetsaolinglandslidedammedlake tion MethodofFloodDuetoDam-breakandItsAp- [J].Journalof Hydrology,2009,374:43-55 plications[j].journalof Sichuan University, [5] ZhanDaojiang,YeShouze.Engineering Hydrolo- 2008,40(1):21-26(,. gy [M].3rdEdition.Beijing:China WaterPower [J].,2008,40 Press,2000(,. ( ) (1):21-26) [M]. :,2000) [13]Peng M,Zhang L M.Analysisof Human Risks [6] Yang Shenghsueh,Pan Yiwen,DongJiajyun.A DuetoDam BreakFloods(Part2):Applicationto SystematicApproachfortheAssessmentofFlood- Tangjiashan Landslide Dam Failure[J].Natural ing Hazardand Risk Associated witha Landslide Hazards,2012,64:

8 AComparativeStudyoftheRapidEstimationofDownstream FloodRiskInducedbyDammedLakes CAOBo 1,2 KANG Ling 1 TAN Debao 2 YANG Shengmei 2 1 ColegeofHydropowerandInformationEngineering,HuazhongUniversityof ScienceandTechnology,Wuhan430074,China 2 ChangjiangRiverScientificResearchInstitute,Changjiang WaterResourcesCommission,Wuhan430010,China Firstauthor:CAOBo,PhDcandidate,specializesinspatialinformationscienceandtechnique. sunbathing@163.com mentalfund,no.cksf /kj;the WuhanScienceandTechnologyProject, No 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 ( 326 ) Abstract:Emergencydisposalofwaterinearthquake-induceddammedlakes mustadequatelyutilize fundamentaldataforeficientdecision making.thispaperpresentsacomparativestudyoftherapid estimationofdownstreamfloodrisksinducedbydammedlakesusingthe Muskingumand Hydrody- namicsmethodscombined withremotesensing,dem,hydrologicdata,and GIStechniques.Both methodswereusedtocalculatethedownstream potentialimpactareafromthetangjiashandammed lakewhenthedam-breakwaterlevelreached690m.theresultsfrom both methodsshowedthatthe downstream BeichuanTownwouldbesubmerged..Therapidity,accuracy,andcomplementaritiesof bothmethodswereevaluatedbasedonthethepotentialimpactarearesults,datarequirements,and calculationtimes. Keywords:dammedlakes;floodrisk;Muskingum method;hydrodynamics;gis Foundationsupport:TheNationalNaturalScienceFoundationofChina,Nos , ;theCenterResearchInstitutesFunda- ofwhichhaveadvantagesanddisadvantagesforfacerecognition.inthispaper,weproposeagaussian weightsbasednew featurecombined withtexturefeatureandstructurefeatureapproach.inthis method,wefirstuseastandardfiducialpointdetectortolocatingfivefacekey-points(e.g.,themid- pointoftheeye),andthenanewfeaturewilbegenerateddynamicalybyweightingthegrayvalueof thepixelaccordingtothedistancebetweenthepixelanditsrelativekey-point.thenewfeature,orig- inatedfromthetexturefeature,tiesthegeometricstructuralfeatureofthepixelsandthekey-pointin- formationanddeliversbeterperformancethantheoriginaltexturefeatureforposevariations.mean- while,thefivekey-pointsarebeneficialtopart-basedfacerecognition.experimentalresultsshowthat thenewfeaturecanimproverecognitionratesbyabout5% thantheoriginaltexturefeature,andim- provedthepart-basedmethodbycombiningthenewfeatureandthefaceblocking;achievingrecogni- tionratescloseto100% fortwodiferentavailabledatabases. Keywords:facerecognition;weightedfeature;faceblocking;LDA,sparserepresentation Firstauthor:FANG Correspondingauthor:LIAO Haibin,PhD,lecturer. liao haibing@163.com Foundationsupport:TheNationalNaturalScienceFoundationofChina,No ;theMajorProgramsforScienceandTechnology DevelopmentofZhengzhouCity,No.072SGZS38042.

:, : ,(4) ,32 21 :,,( ) : ,(2).--3 7

:, : ,(4) ,32 21 :,,( ) : ,(2).--3 7 2006 5 82 10 *****************************************************************...1 21...1...2 10...2...2...3...3...4...4...4...5...5...6...6...7...7...7...8...8...8...9...9...10...10... 11... 11...12...12...13...13