软件导刊2025，Vol.24Issue(5)：194-199,6.DOI:10.11907/rjdk.232256
基于卷积神经网络的地图符号空间冲突检测方法
Detection Method for Cartographic Symbol Space Conflict Based on Convolution Neural Network
摘要
Abstract
The mapping synthesis operations caused by changes in map scale can easily lead to symbol space conflicts,which can be classi-fied into three types:line line conflicts,line surface conflicts,and surface surface conflicts.Detecting conflict types can ensure that the map can still effectively transmit spatial information even after scale changes.Therefore,a convolutional neural network model suitable for symbol space conflict detection is designed to classify symbol space conflicts in small-scale map images with a resolution of 150×150 through data augmentation and equalization methods.Firstly,transform the map image sets of three types of conflicts and no conflicts into a floating-point tensor dataset,and use it as the object of data augmentation and equalization to generate a large-scale tensor dataset;Secondly,the large-scale tensor dataset is fed into a convolutional neural network consisting of one input layer,four pooling layers,three convolutional layers,one flattening layer,one fully connected layer,and one output layer for training,resulting in a network model suitable for symbol space conflict de-tection;Finally,input the map image to be classified into the trained network model to obtain the symbol space conflict type of the image.By using actual map data to verify the performance of the model,it was found that compared with traditional detection methods,the proposed method has higher detection accuracy and has the potential to be applied in practical detection scenarios.关键词
卷积神经网络/符号空间冲突/地图图像/数据增强与均衡Key words
convolution neural network/symbol space conflict/cartographic image/data enhancement and balance分类
计算机与自动化引用本文复制引用
戴涛,蔡武宜,秦晓莉,范琰,刘伟峰..基于卷积神经网络的地图符号空间冲突检测方法[J].软件导刊,2025,24(5):194-199,6.基金项目
国家重点研发计划项目(2019YFE0122600) （2019YFE0122600）
湖南省科技计划项目(2017TP1029) （2017TP1029）
湖南省重点领域研发计划项目(2019SK2101) （2019SK2101）
