基于神经网络与元胞自动机的城市扩展模拟

31120091武汉理工大学学报JOURNALOFWUHANUNIVERSITYOFTECHNOLOGYVol.31No.1Jan.2009DOI:10.3963/j.issn.16714431.2009.01.023詹云军1,2,黄解军2,吴艳艳2(1.,430074;2.,430070):为了探究城市扩展的规律,为城市的规划做出前瞻性的预测,将神经网络与元胞自动机相结合,从不同时相遥感数据中挖掘城市扩展土地利用演变的规律,自动找到土地利用元胞的转换规则,并以该规则反演和预测城市的扩展演变应用该方法对义乌市的扩展作了实证分析和模拟预测,与同期义乌城市发展状况基本相吻合:神经网络;元胞自动机;遥感;模拟;城市扩展:TP183:A:16714431(2009)01008605UrbanExpansionSimulationBasedonArtificialNeuralNetworkandCellularAutomataZHANYunjun1,2,HUANGJiejun2,WUYanyan2(1.InstituteofMathematicalGeologyandRemoteSensingGeology,ChinaUniversityofGeosciences,Wuhan430074,China;2.SchoolofResourcesandEnvironmentEngineering,WuhanUniversityofTechnology,Wuhan430070,China)Abstract:Inordertoexploretherulesofurbanexpansionandtoserveurbanplanning,thispaperresearchedonthecombinationofcellularautomatawithartificialneuralnetwork,andthensetuptheANNCAmode.Themodewasappliedtoexcavatetheruleoftheevolutionoflandusefromdifferenttemporalremotesensingimageandtosimulateurbanexpansion.Inthispaper,theapplicationexamplewasYiwucity.TheresultofsimulatewasconsiderablycoincidentwithsimultaneousexpansioninresearchareainYiwu.Keywords:artificialneuralnetwork;cellularautomata;remotesensing;simulation;urbanexpansion:20080821.:(40572166;40601076).:(1976),,.Email:[e-mail],,,,,,[1](CellularAutomata,CA),CA[2,3],CA[48]CA,,[9,10](ArtificialNeuralNetwork,ANN),ANNCA(ANNCA),,,,1ANNCA1.1CA,,5St+1=(Ud,St,Nt,f)(1),St+1t+1,Udd,,Stt,Ntt,fCA,Ud(d=2),CA1.2,[11],[12][13]3,,,,m,n,,,,,,[0,1]Si(k,t)=[Si(k,t)-Smin]/(SmaS-Smin)(2),SmaS,Smin,Si(k,t)kti1.3ANNCA,CAANNCA[8],,1,;2;3,,,n()1n,tS(k,t)=[S1(k,t),S2(k,t),S3(k,t),!,Sn(k,t)]T(3),T,jnetj(k,t)=∀iWi,jSi(k,t)(4),netj(k,t)j,Wi,j,11+e-netj(k,t)(5),p(k,t,l)=∀jWj,l11+e-netj(k,t)(6),P(k,t,l)ktl;Wj,l,CA[4,8]RA=1+(-ln)(7),[0,1];(6)p(k,t,l)=RA#∀jWi,j11+e-netj(k,t)=[1+(-ln)]#∀jWi,j11+e-netj(k,t)(8),N,,,[0,1],,2ANNCA2.1TM,,87311,:[14,15],,4:,;40([16],),:t,t+1,St∃{,,},St+1=St2.28,8,3,2n/3(n)[17],6,44,1,[2,4,8],,;,1RS2.3,,,,CA2[18,19],(x,y),2,RS,,0,1S1S2S3S4S5S6S7S83,,,,3.119992001200330mTM,,199920012,,,,1000,5002ERDASIMAGE,ThinksPro,,,,(2)328,0,88200913.21999TM,ANNCA2003(3(b))(2003,3(a)),,ANNCA20102020(4),4,,,,,,,,3.322003,87.41%,,2%87.9686.3278.5395.34(),,,,,,2020201033,2010200345655,41km22015(4(b)),3201020032003()57864335663574169762010()1035192898535896697645655-4581015502000%2020(4(c))(),23,,ANNCA,89311,:4a.,,,ANNCA,CA,,,b.,,CA,CA[1],,.%[J].,2002,14(2):4146.[2]WuF,WebsterCJ.SimulationofLandDevelopmentThroughtheIntegrationofCellularAutomataandMulticriteriaEvaluation[J].EnvironmentandPlanningB:PlanningandDesign,1998,25:103126.[3]ClarkeKC,GaydosLJ.LoosecouplingaCellularAutomataModelandGIS:LongtermUrbanGrowthPredictionforSanFranciscoandWashington[J].BaltimoreInternationalJournalofGeographicalInformationScience,1998,12(7):699714.[4]LiX,YehAGO.ModelingSustainableUrbanDevelopmentbytheIntegrationofConstrainedCellularAutomataandGIS[J].InternationalJournalofGeographicalInformationScience,2000,14(2):131152.[5],.CA[J].,1999,54(4):289298.[6]LiX,YehAGO.ZoningforAgriculturalLandProtectionbytheIntegrationofRemoteSensing[J].GISandCellularAutomataPhotogramMetricEngineering&RemoteSensing,2001,67(4):471477.[7],.CellularAutomata[J].,2001,31(8):683690.[8],.[J].,2005,24(1):1927.[9],.[J].:D,2004,34(9):865872.[10],.Fisher[J].,2007,36(1):112118.[11],.Matlab[J].,2006,28(1):108111.[12].Hopfield[J].,2005,29(6):970973.[13].BP[J].,2007(10):2931,53.[14],,.[J].,1999,24(2):166169.[15].CAGIS[J].,1999(11):79.[16],.GIS[J].,2001,30(2):148155.[17]WangF.TheUseofArtificialNeuralNetworksinaGeographicalInformationSystemforAgriculturalLandsuitabilityAssessment[J].EnvironmentandPlanningA,1994,26:265284.[18],.[J].,1997,1(4):282289.[19]WuF.CalibrationofStochasticCellularAutomata:TheApplicationtoRuralurbanLandConversions[J].InternationalJournalofGeographicalInformationScience,2002,16(8):795818.9020091