Google cars generate street views, but not exactly 3D model.
anyone is familiar with equipments/algo to generate interior 3D model or exterior 3D model? Kinect?
It seems UC Berkeley has substantial work on this.
A Photogrammetry solution like Agisoft Photoscan can be used for such tasks.
http://imasgal.com/fotogrametria_escaneado_muralla_romana_lugo/
Agisoft is a good software but it is not specialized for interior scene reconstruction. Some algorithms use stronger assumptions to reconstruct rooms from point clouds base for example on planar walls, etc.
Look at the following:
- http://cg.cs.uni-bonn.de/aigaion2root/attachments/ochmann-2016.pdf
- http://panocontext.cs.princeton.edu/paper.pdf
Folowing th remark from Bruno I can add a nice work :
https://halshs.archives-ouvertes.fr/halshs-00281219/document
A lot of work are now done using a priori knowledge in computer vision and photogrammetry in order to compute camera pose and 3D model.
We can found also similar work to compute camera calibration on facade for example.
- Badges
- Science topic