项目名称: 基于众源影像的三维街景模型增强
项目编号: No.41271431
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 天文学、地球科学
项目作者: 单杰
作者单位: 武汉大学
项目金额: 80万元
中文摘要: 众源影像指由大众使用智能手机或普通数码相机拍摄并通过互联网共享的具有地理坐标的数字影像,其中包含大量时效性好、位置灵活的街景信息。本项目研究利用众源影像对由地面移动测量建立的三维街景模型进行补充和扩展的理论与方法。研究内容包括:1)众源影像的地理配准,即利用已有的街景数据和其它辅助信息确定众源影像的空间位置和姿态;2)众源影像的三维建模,即通过计算机视觉和摄影测量的方法对街景建筑物进行三维重建;3)基于众源影像的街景数据纹理修复,即利用众源影像修复被遮挡的街景数据并对其进行晕光处理,从而形成完整无缝的街景纹理。本项目旨在为众源影像这一新兴数据源在街景建模中的应用奠定理论基础,使街景模型的建立和更新拥有现势和开放的数据源,促进大众参与数字城市建设以及地理学科的大众化发展。
中文关键词: 三维建模;建筑物提取;多角度遥感;众源影像;
英文摘要: Public images or crowd-sourcing images are those with geographic reference, collected by the general public using smart phones or ordinary digital cameras and shared on the Internet. They provide vast, up-to-date, and location flexible data of street views. This research will establish the basic theory and develop the methods for utilizing the public images to enhance and supplement existing street view database created usually by mobile mapping techniques. It includes: 1) geo-referencing of public images to images or models in street view database; 2) 3D building reconstruction from public images by using techniques in computer vision and photogrammetry; and 3) Repairing and replacing occluded textures in existing street view database to generate complete and seamless streetscape textures. This study will lay a theoretic foundation for wide-spreading use of public images - a free, volunteered, up-to-date data source, and promote public involvement in creation of digital city and public awareness of geographic science.
英文关键词: 3D reconstruction;building extraction;multi view sensing;public images;