3DNOW: IMAGE-BASED 3D RECONSTRUCTION AND MODELING VIA WEB
- 1Technologies of Vision - Fondazione Bruno Kessler - Povo (TN), Italy
- 23D Optical Metrology - Fondazione Bruno Kessler - Povo (TN), Italy
- 3LSIS, I&M Team, Aix-Marseille Universite, Polytech Luminy, Marseille, France
- 4Theoretical Physics, ETH Zurich, Zurich, Switzerland
Keywords: Photogrammetry, Structure-from-Motion, Dense Image Matching, 3D modeling, Web service
Abstract. This paper presents a web-based 3D imaging pipeline, namely 3Dnow, that can be used by anyone without the need of installing additional software other than a browser. By uploading a set of images through the web interface, 3Dnow can generate sparse and dense point clouds as well as mesh models. 3D reconstructed models can be downloaded with standard formats or previewed directly on the web browser through an embedded visualisation interface. In addition to reconstructing objects, 3Dnow offers the possibility to evaluate and georeference point clouds. Reconstruction statistics, such as minimum, maximum and average intersection angles, point redundancy and density can also be accessed. The paper describes all features available in the web service and provides an analysis of the computational performance using servers with different GPU configurations.