袁林辉,刘凯,刘佩,沈婕,马劲松

• 1:南京大学地理信息科学系
• 2:南京师范大学虚拟地理环境教育部重点实验室
• 3:南京师范大学地理科学学院
• 4:江苏省地理信息资源开发与利用协同创新中心

摘要(Abstract)：

针对小比例尺道路网专题数据少、语义信息缺乏的特点,该文根据小比例尺道路网已有的语义几何属性和拓扑结构,构建一系列参数作为选取依据。将道路网选取抽象为分类问题,提出一种使用径向基函数神经网络和多参数进行道路网选取的方法。相比时下较常用的反向传播神经网络,径向基函数神经网络具有更高效的学习性能。利用径向基函数神经网络的非线性映射能力,对样本进行训练和验证,并选择不同结构类型(放射式、格网式和自由式)的道路网进行选取实验。结果表明,该方法在小比例尺道路网选取研究中具有可行性,且在自动获取选取参数的同时提高了选取精度。

关键词(KeyWords)： 道路网;智能选取;制图综合;径向基函数;神经网络

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金项目(41371433,41371365)

作者(Author): 袁林辉,刘凯,刘佩,沈婕,马劲松

DOI: 10.16251/j.cnki.1009-2307.2019.03.002

参考文献(References)：

• [1]陈波,武芳,钱海忠.道路网自动选取方法研究[J].中国图象图形学报,2008,13(12):2388-2393.(CHENBo,WU Fang,QIAN Haizhong.Study on road networksauto-selection algorithms[J].Journal of Image andGraphics,2008,13(12):2388-2393.)
• [2]LI Z,ZHOU Q.Integration of linear and arealhierarchies for continuous multi-scale representation ofroad networks[J].International Journal of GeographicalInformation Science,2012,26(5):855-880.
• [3]田晶,郭庆胜,冯科,等.基于信息损失的街道渐进式选取方法[J].武汉大学学报(信息科学版),2009,34(3):362-365.(TIAN Jing,GUO Qingsheng,FENG Ke,et al.Progressive selection approach of streets based oninformation loss[J].Geomatics and Information Scienceof Wuhan University,2009,34(3):362-365.)
• [4]董卫华,郭庆胜,刘纪平,等.道路网适宜性地图的渐进式综合研究[J].武汉大学学报(信息科学版),2007,32(9):829-832,837.(DONG Weihua,GUO Qingsheng,LIU Jiping,et al.Progressive generalization research ofschematic road network maps[J].Geomatics and Infor-mation Science of Wuhan University,2007,32(9):829-832,837.)
• [5]郭敏,钱海忠,黄智深,等.道路网智能选取的案例类比推理法[J].测绘学报,2014,43(7):761-770.(GUOMin,QIAN Haizhong, HUANG Zhishen,et al.Intelligent road-network selection using cases basedreasoning[J].Acta Geodaetica et Cartographica Sinica,2014,43(7):761-770.)
• [6]王家耀.关于信息时代地图学的再思考[J].测绘科学技术学报,2013(4):329-333.(WANG Jiayao.Re-thinkinginformation cartography[J].Journal of Geomatics Scienceand Technology,2013(4):329-333.)
• [7]王家耀,崔铁军,王光霞.图论在道路网自动选取中的应用[J].测绘科学技术学报,1985(1):82-89.(WANGJiayao,CUI Tiejun,WANG Guangxia.Application ofgraph theory in road network selection[J].Journal ofGeomatics Science and Technology,1985(1):82-89.)
• [8]THOMSON R C,RICHARDSON D E.The ‘goodcontinuation’ principle of perceptual organizationapplied to generalization of road networks[C]∥InProceedings of the ICA 19th International CartographicConference.Ottawa:[s.n.],1999.
• [9]JIANG B,CLARAMUNT C.A structural approach tothe model generalization of an urban street network[J].Geoinformatica,2004,8(2):157-171.
• [10]JIANG B,HARRIE L.Selection of streets from anetwork using self-organizing maps[J].Transactions inGIS,2004,8(3):335-350.
• [11]张朋东,邓敏,赵玲,等.集成不同类型特征的城市道路选取方法研究[J].地理与地理信息科学,2011,27(5):16-20.(ZHANG Pengdong,DENG Min,ZHAO Ling,et al.An integrated method of the selection of urbanstreets based on different kinds of properties[J].Geographyand Geo-Information Science,2011,27(5):16-20.)
• [12]蔡永香,郭庆胜.基于神经网络的街道渐进式选取[J].测绘信息与工程.2008,33(5):24-26.(CAI Yong-xiang,GUO Qingsheng.Method for streets progressiveselection based on neural network techniques[J].Journalof Geomatics,2008,33(5):24-26.)
• [13]刘凯,李进,沈婕,等.基于BP神经网络和拓扑参数的道路网选取研究[J].测绘科学技术学报,2016(3):325-330.(LIU Kai,LI Jin,SHEN Jie,et al.Selection ofroad network using bp neural network and topologicalparameters[J].Journal of Geomatics Science andTechnology,2016(3):325-330.)
• [14]POWELL M J D.Radial basis functions formultivariable interpolation:a review[J].Algorithms forApproximation,1987:143-167.
• [15]POGGIO T,GIROSI F.Regularization algorithms forlearning that are equivalent to multilayer networks[J].Science,1990,247(4945):978-82.
• [16]POGGIO T,GIROSI F.Networks for approximationand learning[J].Proceedings of the IEEE,1990,78(9):1481-1497.
• [17]LOOMER C W.Elements of cartography[M]∥Elements of cartography.Wiley:[s.n.],1995:80-82.
• [18]THOMSON R C.The’stroke’concept in geographicnetwork generalization and analysis[M].Progress inSpatial Data Handling.Berlin Heidelberg:Springer,2006:681-697.
• [19]LI Z,CHOI Y H.Topographic map generalization:association of road elimination with thematic attributes[J].Cartographic Journal,2002,39(2):153-166.
• [20]HU Yungang,CHEN Jun,LI Zhilin,et al.Selectiveomission of road features based on mesh density fordigital map generalization[J].Acta Geodaetica etCartographica Sinica,2007,36(3):351-357.
• [21]陈波,武芳,钱海忠.空间分析方法在道路网自动综合拓扑构建中的应用[J].测绘科学,2006,31(1):101-102.(CHEN Bo,WU Fang,QIAN Haizhong.Appli-cation of spatial analysis in topology constructing ofroad network generalization[J].Science of Surveyingand Mapping,2006,31(1):101-102.)
• [22]TOUYA G.A road network selection process based ondata enrichment and structure detection[J].Tran-sactions in GIS,2010,14(5):595-614.
• [23]杨敏,艾廷华,周启.顾及道路目标stroke特征保持的路网自动综合方法[J].测绘学报,2013,42(4):581-587.(YANG Min,AI Tinghua,ZHOU Qi.A method ofroad network generalization considering strokeproperties of road object[J].Acta Geodaetica etCartographica Sinica,2013,42(4):581-587.)
• [24]MOODY J,DARKEN C J.Fast learning in networks oflocally-tuned processing units[J].Neural Computation,1989,1(2):281-294.
• [25]王玲,龚健雅.基于RBF神经网络的遥感影像分类算法研究[J].测绘工程,2003,12(3):7-9.(WANG Ling,GONG Jianya.The research on image classificationalgorithm based on the RBF neural network[J].Engi-neering of Surveying and Mapping,2003,12(3):7-9.)
• [26]NIROS A D,TSEKOURAS G E.A novel trainingalgorithm for RBF neural network using a hybrid fuzzyclustering approach[J].Fuzzy Sets &Systems,2012,193(9):62-84.
• [27]ZHANG J,WANG Y,ZHAO W.An improved hybridmethod for enhanced road feature selection in mapgeneralization[J].International Journal of Geo-Information,2017,6(7):196.
• [28]BENZ S A,WEIBEL R.Road network selection formedium scales using an extended stroke-mesh combinationalgorithm[J].Cartography and Geographic InformationScience,2014,41(4):323-339.