利用经典CNN网络方法构建贵阳市道路要素遥感影像自动提取模型  

Using the classical CNN network method to construct the automatic extraction model of remote sensing image of Guiyang road elements

作　　者：佘佐明 申勇智 宋剑虹 向玉瑨 SHE Zuoming;SHEN Yongzhi;SONG Jianhong;XIANG Yujin(Guiyang Institute of Surveying and Mapping,Guiyang 550000,China)

机构地区：[1]贵阳市测绘院,贵州贵阳550000

出　　处：《测绘通报》2023年第4期177-182,共6页Bulletin of Surveying and Mapping

摘　　要：综合考虑道路提取解译过程中的准确性、运算能力及对贵阳市域环境的适应性,本文对深度学习神经网络模型中的几个环节进行了分解,通过多轮对比试验与分析,建立了适用于贵阳市道路要素遥感影像自动提取的模型,并对批量提取的数据进行分析和优化处理,完成部分道路属性的填充,较大程度地实现了道路实体的自动化智能高效提取。过程中涉及的现实问题与技术路线,可对市县级卫星遥感应用技术部门开展的自然资源类业务工作提供参考。Road extraction comprehensively consider the accuracy,computing ability and adaptability to the environment of Guiyang in the interpretation process,so several links in the deep learning neural network model are decomposed.Through multiple rounds of comparative experiments and analysis,a model for automatic extraction of remote sensing images of road elements in Guiyang is established in this paper,the data extracted in batches are analyzed and optimized to complete the filling of some road attributes,which largely realizes the automatic intelligent and efficient extraction of road entities.The practical problems and technical routes involved in the process can provide reference for the natural resources business carried out by the municipal and county level satellite remote sensing application technology departments.

关 键 词：CNN 深度学习 道路提取 遥感解译 

分 类 号：P237[天文地球—摄影测量与遥感]