Volume XL-4/W5
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XL-4/W5, 137-141, 2015
https://doi.org/10.5194/isprsarchives-XL-4-W5-137-2015
© Author(s) 2015. This work is distributed under
the Creative Commons Attribution 3.0 License.
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XL-4/W5, 137-141, 2015
https://doi.org/10.5194/isprsarchives-XL-4-W5-137-2015
© Author(s) 2015. This work is distributed under
the Creative Commons Attribution 3.0 License.

  11 May 2015

11 May 2015

CREATING PRODUCT MODELS FROM POINT CLOUD OF CIVIL STRUCTURES BASED ON GEOMETRIC SIMILARITY

N. Hidaka1, T. Michikawa2,3, N. Yabuki1, T. Fukuda1, and A. Motamedi1 N. Hidaka et al.
  • 1Division of Sustainable Energy and Environmental Engineering, Osaka University, Suita, Osaka, Japan
  • 2Center for Environmental Innovation Design for Sustainability, Osaka University, Suita, Osaka, Japan
  • 3Photonics Control Technology Team, RIKEN, Wako, Saitama, Japan

Keywords: Point Cloud, CIM, Geometric Similarity, Surface Reconstruction

Abstract. The existing civil structures must be maintained in order to ensure their expected lifelong serviceability. Careful rehabilitation and maintenance planning plays a significant role in that effort. Recently, construction information modelling (CIM) techniques, such as product models, are increasingly being used to facilitate structure maintenance. Using this methodology, laser scanning systems can provide point cloud data that are used to produce highly accurate and dense representations of civil structures. However, while numerous methods for creating a single surface exist, part decomposition is required in order to create product models consisting of more than one part. This research aims at the development of a surface reconstruction system that utilizes point cloud data efficiently in order to create complete product models. The research proposes using the application of local shape matching to the input point clouds in order to define a set of representative parts. These representative parts are then polygonized and copied to locations where the same types of parts exist. The results of our experiments show that the proposed method can efficiently create product models using input point cloud data.