Voxel Carving for Specular Surfaces

  • Authors:

  • Thomas Bonfort;Peter Sturm

  • Affiliations:

  • -;-

  • Venue:
  • ICCV '03 Proceedings of the Ninth IEEE International Conference on Computer Vision - Volume 2
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present an novel algorithm that reconstructs voxels of a general3D specular surface from multiple images of a calibrated camera. Acalibrated scene (i.e. points whose 3D coordinates are known) isreflected by the unknown specular surface onto the image plane ofthe camera. For every viewpoint, surface normals are associated tothe voxels traversed by each projection ray formed by thereflection of a scene point. A decision process then discardsvoxels whose associated surface normals are not consistent with oneanother. The output of the algorithm is a collection of voxels andsurface normals in 3D space, whose quality and size depend onuser-set thresholds. The method has been tested on synthetic andreal images. Visual and quantified experimental results arepresented.