基于生成式对抗网络的地形特征线提取方法  

Study on Extraction Method of Terrain Features Lines Based on Generative Adversarial Networks

作　　者：魏子婷 杜清运[1,2] 王沛文 WEI Ziting;DU Qingyun;WANG Peiwen(School of Resources and Environmental Sciences,Wuhan University,Wuhan 430079,China;Key Laboratory of Geographic Information System,Ministry of Education,Wuhan University,Wuhan 430079,China;StarGIS(Tianjin)Technology Development Co.,Ltd,Tianjin 30084,China)

机构地区：[1]武汉大学资源与环境科学学院,湖北武汉430079 [2]武汉大学地理信息系统教育部重点实验室,湖北武汉430079 [3]星际空间(天津)科技发展有限公司,天津30084

出　　处：《地理信息世界》2021年第3期20-24,共5页Geomatics World

基　　金：国家重点研发计划项目(2016YFC0803106)。

摘　　要：地形特征线反映了地形表面的起伏形态,其提取对地理分析和研究具有重大意义。近年来,深度学习为图像处理技术提供了巨大的发展潜力。研究提出了基于深度学习模型——pix2pix的地形特征线提取方法,在山地和高原两种地形下进行了实验,并且将该方法应用于多比例尺底图的地形等高线提取研究。实验结果表明,pix2pix模型不仅在地形特征线提取方面有良好的效果,并且在特定的山地地形下可以进行山脊线和山谷线的语义区分。Terrain features lines reflect the characteristic of topographic surface,and their extraction is of great significance for geographical analysis and research.In recent years,deep learning has great potential for image processing technology.In this study,an extraction method of terrain features lines is proposed based on a deep learning model pix2pix.Experiments are conducted in mountain terrains and plateau terrains of multi-scale maps.Results show that pix2pix model not only has good effect in extraction of terrain features lines,but also can distinguish ridge lines and valley lines in specific mountain terrains.

关 键 词：地形特征线 生成式对抗网络 卷积神经网络 深度学习 

分 类 号：P208[天文地球—地图制图学与地理信息工程]