高分辨率遥感影像道路提取方法研究进展

Advances in Method on Road Extraction from High Resolution Remote Sensing Images

快速、准确地从高分辨率遥感影像中提取道路信息对于基础地理信息更新具有重要的意义。针对道路提取的众多方法与技术,首先对遥感影像道路提取中涉及的特征类型与表达方法进行了归纳;其次,参照提取层次将道路提取方法分为3种类型:基于特征、基于对象以及基于知识的道路提取,并对每一类型道路提取方法的优缺点进行了系统总结,同时也对目前高分影像道路提取中存在的问题进行了分析;最后,结合遥感与模式识别的最新发展等相关理论,对高分影像道路提取方法的发展方向进行了展望。

Automatic road extraction quickly and precisely from high resolution remote sensing imagery has huge implication for fundamental geographic information upgrading.based on the existing road extrac- tion technological approaches.This thesis,firstly, made a minutely generalization for feature expression and extraction involved in road extraction approaches.And secondly, the present methods was categorized into three types： feature-based, object-based and knowledge-based. The advantages and disadvantages of each type was systematically summarized separately.In addition,the existential issues was analyzed in detail.Fi- nally,the development trends of road extraction from high resolution remote sensing imagery were pros- pected combining with the development direction of remote sensing and pattern recognition.