基于高分辨率遥感影像的建筑物提取  

Buildings extraction based on high-resolution remote sensing imagery

作　　者：王丽梅 王延正 WANG Limei;WANG Yanzheng(Hydrology Engineering Geological Exploration Institute of Hebei Province,Shijiazhuang 050021,China;Hebei Provincial Remote Sensing Center,Shijiazhuang 050021,China;Beijing MapCore Technology Co.,Ltd.,Beijing 100192,China)

机构地区：[1]河北省水文工程地质勘查院(河北省遥感中心),河北石家庄050021 [2]河北省遥感中心,河北石家庄050021 [3]北京图源科技有限公司,北京100192

出　　处：《测绘通报》2023年第6期180-183,共4页Bulletin of Surveying and Mapping

摘　　要：高分辨率遥感影像不仅具有丰富的光谱、空间分布、形状和纹理特征,也包含清晰的场景语义信息。本文以安徽省枞阳县枞阳镇为研究区域,以高分辨率影像为基础数据源,利用eCognition软件中深度学习与面向对象相结合的方法进行建筑物自动提取。结果表明,该方法具有更好的建筑物提取效果,总体分类精度达96.8%,可用于通过高分辨率影像进行建筑物提取的生产。High-resolution remote sensing images not only have rich spectrum,spatial distribution,shape and texture features,but also contain clear scene semantic information.Taking Zongyang town,Zongyang county,Anhui province as the research area,and using high-resolution images as the basic data source,the deep learning and object-oriented method in eCognition software is used to automatically extract buildings in this paper.The results show that the method of combining deep learning with object-oriented has a better effect of building extraction,and the overall classification accuracy reaches 96.8%,which can be used for building extraction production based on high-resolution images.

关 键 词：深度学习 ECOGNITION 多尺度分割 面向对象影像分析 卷积神经网络 

分 类 号：P237[天文地球—摄影测量与遥感]