講演情報
[D3-03]Reconstructing 3d Segmentation of the Urban Environment using Street View Images
*Monykandy Leng1, Zongchao Gu2, Jun Cao3 (1. Master Student, School of Architecture, Southeast University, 2. Associate Professor, School of Architecture, Southeast University, 3. Lecturer, School of Architecture, Southeast University)
キーワード：
3d segmentation、3d reconstruction、panoramic images、street view images
The recreation of the environment can be a complex process involving a lot of resources and time to produce. Even if the city has been produced digitally, the maintaining and updating process would require continuous investments. The study proposes a method to produce the 3-dimensional spatial relations using only available street view images as the source of the input. First, the images are pre-process before producing the 3d environments, The resulted point cloud is later combined with segmentation of the images producing a the segmented 3d environments. This results not only the recreation of the environments but its segmentation can also be used in other urban study purposes, such as seismic risk assessment, simulations, or data recreation for mid to small scale study of the urban area. The researches introduce a tool to easily obtain the spatial relations of the city using minimal resources and can be used for other urban study purposes.