Applying pose estimation techniques with structure from motion in architectural fields

SfM (Structure from Motion) is a kind computer vision that creates a point cloud of surrounding scenery and camera poses from photographs Our previous studies have shown that camera poses and their convenient estimation technique can provide beneficial uses in architectural fields. This paper reviews the availability of SfM as a pose estimation technique. The availability is studied in two uses: photographic measurement and location-based photograph management. In photographic measurement, camera poses are adequately estimated and measuring results show applicative accuracy. Location-based photograph management proposes a way to plot 3D CAD and a searchable database of new photographs? poses. © 2018 Architectural Institute of Japan. All rights reserved.