基于MBR组合优化算法的多尺度面实体匹配方法

第４７卷第５期测绘学报Vol．４７,No．５２０１８年５月ActaGeodaeticaetCartographicaSinicaMay,２０１８引文格式:刘凌佳,朱道也,朱欣焰,等．基于MBR组合优化算法的多尺度面实体匹配方法[J]．测绘学报,２０１８,４７(５):６５２Ｇ６６２．DOI:１０．１１９４７/j．AGCS．２０１８．２０１６０６２５．LIULingjia,ZHUDaoye,ZHUXinyan,etal．AMultiＧscalePolygonalObjectMatchingMethodBasedonMBRCombinatorialOptimizationAlgorithm[J]．ActaGeodaeticaetCartographicaSinica,２０１８,４７(５):６５２Ｇ６６２．DOI:１０．１１９４７/j．AGCS．２０１８．２０１６０６２５．基于MBR组合优化算法的多尺度面实体匹配方法刘凌佳１,朱道也１,朱欣焰１,２,３,丁小辉４,呙维１,２１．武汉大学测绘遥感信息工程国家重点实验室,湖北武汉４３００７９;２．武汉大学地球空间信息技术协同创新中心,湖北武汉４３００７９;３．武汉大学空天信息安全与可信计算教育部重点实验室,湖北武汉４３００７２;４．中国科学院东北地理与农业生态研究所,吉林长春１３０１０２AMultiＧscalePolygonalObjectMatchingMethodBasedonMBRCombinatorialOptimizationAlgorithmLIULingjia１,ZHUDaoye１,ZHUXinyan１,２,３,DINGXiaohui４,GUOWei１,２１．StateKeyLaboratoryofInformationEngineeringinSurveying,MappingandRemoteSensing,WuhanUniversity,Wuhan４３００７９,China;２．CollaborativeInnovationCenterofGeospatialTechnology,WuhanUniversity,Wuhan４３００７９,China;３．KeyLaboratoryofAerospaceInformationSecurityandTrustedComputingofMinistryofEducation,WuhanUniversity,Wuhan４３００７２,China;４．NortheastInstituteofGeographyandAgroecology,ChineseAcademyofSciences,Changchun１３０１０２,ChinaAbstract:AimingtosolvingtheproblemofpositionaldiscrepancyofcorrespondingobjectsinmultiＧscalepolygonalobjectmatchingandthatthepotentialmatchingpairscan’tbedirectlyidentifiedbythemethodofarealoverlapping,itisproposedthatamultiＧscalepolygonalobjectmatchingmethodbasedonminimumboundingrectanglecombinatorialoptimizationalgorithm．Thebasicideaofourmethodisthat:①identifyingthepotentialmatchingpairsof１∶１,１∶NandM∶Nwithcombinatorialalgorithmandsimpleshapecharacteristic;②establishingmultiＧcharacteristicartificialneuralnetworkmodeltoevaluatethesepotentialmatchingpairs．Theproposedmethodisdemonstratedintheexperimentofmatchingbetween１∶２０００and１∶１００００polygonalobjectsofresidentialbuildingsandindustrialfacilitiesinZhoushan,ZhejiangProvince．Theexperimentalresultsshowedthattheproposedmatchingmethodshowsuperiorperformanceagainstamethodofareaoverlappingandartificialneuralnetwork．Itsp