Update: Code online now: https://github.com/strasdat/ScaViSLAM

This is the project page of ScaViSLAM - a general and scalable framework for visual SLAM. It employs "Double Window Optimization" (DWO) as described in our ICCV publication:

H. Strasdat, A.J. Davison, J.M.M. Montiel, and K. Konolige, Double Window Optimisation for Constant Time Visual SLAM Accepted for the IEEE International Conference on Computer Vision (ICCV), 2011. pdf

This SLAM framework solves various different kind of visual SLAM tasks such as large loopy outdoor SLAM, mapping local loopy motion (e.g. for augmented reality), or creating dense object models. Have a look at the corresponding video:

• This software package is still under development but the sources are now public and available at github:

https://github.com/strasdat/ScaViSLAM

Note, at the moment only sources for stereo/RGB-D cameras are available. Also, there is no ROS wrapper (yet).

Btw. the first sequence in the video is showing the "New College" in Oxford, UK. The data set is taken from the following website: http://www.robots.ox.ac.uk/NewCollegeData/

Hauke Strasdat (http://www.doc.ic.ac.uk/~strasdat/)