Dense 3D Mapping Using Volume Registration from Monocular View
In previous work it has been shown that volume registration can be used to generate dense 3D maps from RGBD video data with various advantages over feature matching approaches. This paper presents preliminary work in applying this technique to general monocular video data. This approach does not require additional RGB-D hardware and generates dense 3D models with similar complexity.
2016 10th International Conference on Complex, Intelligent, and Software Intensive Systems