脉冲耦合神经网络支持下的网络Voronoi图构建  

Algorithm for constructing network Voronoi diagram based on improved pulse coupled neural network

作　　者：禄小敏 闫浩文 李小军 王杭宇 武芳[3] LU Xiaomin;YAN Haowen;LI Xiaojun;WANG Hangyu;WU Fang(Faculty of Geomatics,Lanzhou Jiaotong University,Lanzhou 730070;Gansu Provincial Engineering Laboratory for National Geographic State Monitoring,Lanzhou 730070,China;Faculty of Geomatics,Information Engineering University,Zhengzhou 450000,China)

机构地区：[1]兰州交通大学测绘与地理信息学院,兰州730070 [2]甘肃省地理国情监测工程实验室,兰州730070 [3]信息工程大学地理空间信息学院,郑州450000

出　　处：《测绘科学》2021年第3期169-175,共7页Science of Surveying and Mapping

基　　金：国家自然科学基金重点项目(41930101);国家自然科学基金青年项目(41801395);国家自然科学基金地区项目(41671447)。

摘　　要：针对网络Voronoi图中点与点之间通过实际路径距离而非传统欧式距离相连,在实际应用中较平面Voronoi图更加合理,该文提出了一种基于改进脉冲耦合神经网络的网络Voronoi图构建算法。借助模型的自动波发放及并行处理特性,较好地实现了基于路网的网络空间剖分,顾及了道路网及其点群自身属性对其服务范围的影响。实验表明,该算法实现了点群网络Voronoi图的构建,最短路径思想的引入使得构建的网络Voronoi图符合Voronoi图基本特征,可以用来表示点群的服务范围,不仅如此,算法的并行特性保证了算法的高效率。Network Voronoi diagram is more reasonable in real application than traditional planar Voronoi diagram,because the connection between two points in network Voronoi diagram is represented as real network distance instead of Euclidean distance.Aiming at establishing the network Voronoi diagram model,a construction algorithm based on improved PCNN is proposed.With the characteristics of the automatic wave emission and parallel processing,the network space partition is well realized,and the influence of road network and point cluster on the service scope is considered.Experiments show that the algorithm can construct network Voronoi diagram well,the constructed network Voronoi diagram imported the idea of the minimum distance match the character of Voronoi diagram,and it can be used to represent service regions of the points.In addition,the algorithm is high in efficiency dues to its concurrent feature.

关 键 词：网络Voronoi图 改进PCNN模型 神经元 最短路径 

分 类 号：P208[天文地球—地图制图学与地理信息工程]