论文部分内容阅读
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 the focus of remote sensing image automatic / semi-automatic road extraction. This paper first introduces the traditional video road model and points out the shortcomings. It also illustrates the necessity of establishing the road semantic model from the perspective of semantics. It also summarizes the performance characteristics of 10m-resolution satellite remote sensing image road, At each level of semantic information, the road semantic model for 10m resolution satellite remote sensing image is established, and the realization process of the semantic model is elaborated. The semantic model is applied to the remote sensing image road extraction system developed by the author. The road semantic model of the road is validated by the Spot5 multispectral remote sensing image road extraction test in Atushi, Xinjiang.