| [1] |
ZhangHao,WangQijie,ZhuJianjun,etal.Influ enceofSampleDataPreprocessingonBP NeuralNetwork basedGPSElevationFtting[J].犑狅狌狉狀犪犾狅犳犌犲狅犱犲狊狔犪狀犱犌犲狅犱狔狀犪犿犻犮狊,2011,31(2):125 128(张昊,王琪洁,朱建军,等.样本数据预处理对基于BP神经网络的GPS高程拟合的影响[J].大地测量与地球动力学,2011,31(2):125 128)[2] MiaoBaotong,ChenFalai.ApplicationofRadiusBasisFunctionNeuralNetworksinScatteredDataInterpolation[J].犑狅狌狉狀犪犾狅犳犆犺犻狀犪犝狀犻狏犲狉狊犻狋狔狅犳犛犮犻犲狀犮犲犪狀犱犜犲犮犺狀狅犾狅犵狔,2001,31(2):135 142(缪报通,陈发来.径向基函数神经网络在散乱数据插值中的应用[J].中国科技技术大学学报,2001,31(2):135 142)[3] SongLei,FangJian,ZhouXuhua,etal.Applica tionof Bayesian Regulation BP Neural NetworkTofitTwoKindQuasigeoid[J].犌犲狅犿犪狋犻犮狊犪狀犱犐狀犳狅狉犿犪狋犻狅狀犛犮犻犲狀犮犲狅犳 犠狌犺犪狀犝狀犻狏犲狉狊犻狋狔,2009,34(5):552 555(宋雷,方剑,周旭华,等.Bayesian正则化BP神经网络拟合两类似大地水准面[J].武汉大学学报·信息科学版,2009,34(5):552 555)[4] ZhangZhenglu,WangXiaomin,DengYong,etal.ApplicationofFuzzyNeuralNetworkinDeforma tionAnalysisandPrediction[J].犌犲狅犿犪狋犻犮狊犪狀犱犐狀犳狅狉犿犪狋犻狅狀犛犮犻犲狀犮犲狅犳 犠狌犺犪狀犝狀犻狏犲狉狊犻狋狔,2010,35(1):7 8(张正禄,王小敏,邓勇,等.模糊神经网络在变形分析与预报中的应用研究[J].武汉大学学报·信息科学版,2010,35(1):7 8)[5] HeHaiqing,HuangShengxiang,WuGen.HeightFittingbyRadialNeuralNetworkfortheConstruc tionQualityControlofFaceRockfillDam[J].犌犲狅犿犪狋犻犮狊犪狀犱犐狀犳狅狉犿犪狋犻狅狀犛犮犻犲狀犮犲狅犳 犠狌犺犪狀犝狀犻狏犲狉狊犻狋狔,2012,37(5):594 597(何海清,黄声享,伍根.528 第40卷第4期 何海清等:一种顾及粗差的径向神经网络高程曲面拟合法碾压施工质量监控的径向神经网络拟合高程研究[J].武汉大学学报·信息科学版,2012,37(5):594 597)[6] MarcekM,MarcekD.GranularRBFNeuralNet workImplementationofFuzzySystems:ApplicationtoTimeSeriesModeling[J].犕狌犾狋犻狆犾犲犞犪犾狌犲犱犔狅犵犻犮犪狀犱犛狅犳狋犆狅犿狆狌狋犻狀犵,2008,14(3/5):401 414[7] TangZesheng.Visualizationof3DDataField[M].Beijing:TsinghuaUniversityPress,1999(唐泽圣.三维数据场可视化[M].北京:清华大学出版社,1999)[8] YangXiaoyun,Guliya,CenMinyi,etal.A Meth odBasedFittingof Moving QuadricSurface withVariableWindowsforDetectingGrossErrorsinIr regularDEM[J].犃犮狋犪犌犲狅犱犪犲狋犻犮犪犲狋犆犪狉狋狅犵狉犪狆犺犻犮犪犛犻狀犻犮犪,2005,34(2):148 153(杨晓云,顾利亚,岑敏仪,等.基于不同大小窗口的移动曲面拟合法探测不规则DEM粗差的一 种方 法 [J].测 绘 学报,2005,34(2):148 153)[9] ShanJie.AnInteractiveMethodforBlunderDetec tioninDEMData[J].犑狅狌狉狀犪犾狅犳犘犔犃犐狀狊狋犻狋狌狋犲狅犳犛狌狉狏犲狔犻狀犵犪狀犱犕犪狆狆犻狀犵,1993,2:28 31(单杰.一种交互式DEM粗差检测方法[J].解放军测绘学院学报,1993,2:28 31)[10]LoweD G.ObjectRecognitionfrom LocalScale InvariantFeatures[C].TheIEEE InternationalConferenceonComputerVision,Kerkyra,Greele,1999犌狉狅狊狊犈狉狉狅狉犆狅狀狊犻犱犲狉犲犱犎犲犻犵犺狋犛狌狉犳犪犮犲犉犻狋狋犻狀犵犝狊犻狀犵犚犪犱犻犪犾犖犲狌狉犪犾犖犲狋狑狅狉犽犎犈犎犪犻狇犻狀犵1 犎犝犃犖犌犛犺犲狀犵狓犻犪狀犵2 犆犎犈犖犜犻狀犵31 FacultyofGeomatics,EastChinaInstituteofTechnology,Nanchang330013,China2 SchoolofGeodesyandGeomatics,WuhanUniversity,Wuhan430079,China3 SchoolofWaterResourcesandHydropower,WuhanUniversity,Wuhan430072,China犃犫狊狋狉犪犮狋:Heightsurfacefittingbyradialneuralnetworkissensitivetogrosserroroftrainingsam ples,aapproachofgrosserrorconsideredheightsurfacefittingusingradialneuralnetworkisintro ducedbasedontechniquesofdigitalimageprocessing,allheightgridpointsbyradialneuralnetworkwithgrosserroraremappedtograyimage,imagedifferenceofGaussianandleastsquaremovingcurvedsurfacealgorithmsareimplementedforgrosserrordetectionandcorrection,andimplementingheightfittingbyradialneuralnetworkwithgrosserrorcorrecteddiscreteheightpoints.Toverifyitsfeasibility,weimplementedheightfittingbyrobustradialneuralnetworkwiththedatafromsurvey ingrobot.Theresultsshowthatthemethodcanachievebetteraccuracy,hasstrongrobustnessandfeasibility.犓犲狔狑狅狉犱狊:radialneuralnetwork;grosserrordetection;differenceofGaussian;movingcurvedsur facemethod;heightsurfacefitting犉犻狉狊狋犪狌狋犺狅狉:HEHaiqing,PhD,specializesinthephotogrammetryandremotesensing,geodeticdataprocessing.E mail:hyhqing@163.com犉狅狌狀犱犪狋犻狅狀狊狌狆狆狅狉狋:TheNationalNaturalScienceFoundationofChina,No.41401526;theOpenResearchFundofStateKeyLaboratoryofInformationEngineeringinSurveying,MappingandRemoteSensing,No.(13)04;theOpenResearchFundofKeyLaboratoryofSat elliteMapping Technologyand Application,NationalAdministrationofSurveying,Mappingand Geoinformation,No.KLAMTA 201404.
|
百度学术
下载:
