摘要
道路模型是遥感影像自动/半自动道路提取中重点研究的内容。本文首先介绍了传统的影像道路模型并指出其中不足,从语义的角度阐述了建立道路语义模型的必要性,概述了10m级分辨率卫星遥感影像道路的表现特征,分析了遥感影像道路提取涉及的各层次语义信息,建立了面向10m级分辨率卫星遥感影像的道路语义模型,详细阐述了语义模型的实现过程。将语义模型应用于笔者开发的遥感影像道路提取系统,通过新疆阿图什Spot5多光谱遥感影像道路提取试验,验证了道路语义模型的有效合理性。
Road model is an important content of study on automatic/semi-automatic road extraction from remote sensing images. In the paper, traditional image road models are firstly analysed and their respective shortcomings are summarized. Then from the semantic point of view, the necessary of setting uproad semantic model is elaborated. After that, the paper summaies road characteristics on the level of 10 meter resolution satellite remote sensing images. Having analyzed semantic layers and semantic information of each layer involved in extracting road from remote sensing images, the paper builds a road semantic model mainly for the level of 10 meter resolution satellite images and detailedly describes its process of computer realization. Finally, through an experiment of road extraction from a Spot-5 multi-spectral image of Atush city in Xinjiang Uyghur Autonomous Region under the use of our road extraction system with intelligent copyright, it indicates that this road semantic model is effective.
出处
《测绘科学》
CSCD
北大核心
2009年第3期68-70,202,共4页
Science of Surveying and Mapping
基金
国家863重点项目"多源遥感数据综合处理与服务系统"资助
关键词
卫星影像
道路提取
语义模型
medium-resolution satellite remote sensing images
road extraction
semantic model
