基于RBF神经网络的OSM道路网智能选取  

Intelligent Selection of OSM Road Network Based on RBF Neural Network

作　　者：刘佩 袁林辉 张康 沈婕[2,3,4] 马劲松 LIU Pei;YUAN Linhui;ZHANG Kang;SHEN Jie;MA Jinsong(Department of Geographic Information Science, Nanjing University, Nanjing 210023, China;Key Laboratory of Virtual Geographic Environment of Ministry of Education, Nanjing Normal University,Nanjing 210046, China;Institute of Geographic Science, Nanjing Normal University, Nanjing 210046,China;Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application, Nanjing 210023, China)

机构地区：[1]南京大学地理信息科学系,江苏南京210023 [2]南京师范大学虚拟地理环境教育部重点实验室,江苏南京210046 [3]南京师范大学地理科学学院,江苏南京210046 [4]江苏省地理信息资源开发与利用协同创新中心,江苏南京210023

出　　处：《地理信息世界》2019年第3期8-13,共6页Geomatics World

基　　金：科技部政府间国际科技创新合作重点专项(2016YFE0131600);国家自然科学基金项目(41871371,41371365)资助

摘　　要：针对目前自动制图综合方法在软件平台没有普及,电子地图仍以人工编制为主,Open Street Map(简称OSM)道路不适合作为标准地图等问题,提出了一种将径向基函数神经网络(Radial Basis Function, RBF)应用在OSM道路网自动选取中的方法。根据多种常用的语义、几何、拓扑参数综合考虑道路网的重要性,并将该算法在系统平台中加以实现,以期为OSM道路网自动选取的精度提高和应用普及提供解决方案。实验结果与实际制图结果在形状结构上保持良好,精度为86.92%,相较于BP(Back Propagation)神经网络算法的效果有所提高。Open Street Map (OSM) data is not suitable to be used as a standard map. Considering this, in addition to the situation that automated cartographic generalization methods are not popular in the software platform and the compilation of electronic map is still mainly manually, a radial basis function neural network (RBF) is built to apply in the automatic selection of the OSM road network. Combining self-organization and supervised learning, RBF learns faster than the commonly used back-propagation neural network. Taking into account the importance of the commonly used semantic, geometric, topological parameters of the road networks, an algorithm is developed and then implements in the software platform in order to achieve the solution with high precision selection of the OSM road network. The experiments with OSM road network show that the experimental results and the standard map are in good shape with the accuracy of 86.92%, which is better than BP algorithm.

关 键 词：OSM 径向基函数 神经网络 道路网选取 自动制图综合 

分 类 号：TP393.01[自动化与计算机技术—计算机应用技术]