InfiniTAM v3: a framework for large-scale 3D reconstruction with loop closure
Volumetric models have become a popular representation for 3D scenes in recent years. One breakthrough leading to their popularity was KinectFusion, which focuses on 3D reconstruction using RGB-D sensors. However, monocular SLAM has since also been tackled with very similar approaches. Representing...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Internet publication |
| Language: | English |
| Published: |
ArXiv
2017
|
| _version_ | 1811139696304062464 |
|---|---|
| author | Prisacariu, VA Kähler, O Golodetz, S Sapienza, M Cavallari, T Torr, P Murray, DW |
| author_facet | Prisacariu, VA Kähler, O Golodetz, S Sapienza, M Cavallari, T Torr, P Murray, DW |
| author_sort | Prisacariu, VA |
| collection | OXFORD |
| description | Volumetric models have become a popular representation for 3D scenes in recent years. One breakthrough leading to their popularity was KinectFusion, which focuses on 3D reconstruction using RGB-D sensors. However, monocular SLAM has since also been tackled with very similar approaches. Representing the reconstruction volumetrically as a TSDF leads to most of the simplicity and efficiency that can be achieved with GPU implementations of these systems. However, this representation is memory-intensive and limits applicability to small-scale reconstructions. Several avenues have been explored to overcome this. With the aim of summarizing them and providing for a fast, flexible 3D reconstruction pipeline, we propose a new, unifying framework called InfiniTAM. The idea is that steps like camera tracking, scene representation and integration of new data can easily be replaced and adapted to the user's needs.
This report describes the technical implementation details of InfiniTAM v3, the third version of our InfiniTAM system. We have added various new features, as well as making numerous enhancements to the low-level code that significantly improve our camera tracking performance. The new features that we expect to be of most interest are (i) a robust camera tracking module; (ii) an implementation of Glocker et al.'s keyframe-based random ferns camera relocaliser; (iii) a novel approach to globally-consistent TSDF-based reconstruction, based on dividing the scene into rigid submaps and optimising the relative poses between them; and (iv) an implementation of Keller et al.'s surfel-based reconstruction approach. |
| first_indexed | 2024-09-25T04:10:11Z |
| format | Internet publication |
| id | oxford-uuid:36cf7111-17cd-4c09-a71d-450253d02f64 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-09-25T04:10:11Z |
| publishDate | 2017 |
| publisher | ArXiv |
| record_format | dspace |
| spelling | oxford-uuid:36cf7111-17cd-4c09-a71d-450253d02f642024-06-20T15:47:34ZInfiniTAM v3: a framework for large-scale 3D reconstruction with loop closureInternet publicationhttp://purl.org/coar/resource_type/c_7ad9uuid:36cf7111-17cd-4c09-a71d-450253d02f64EnglishSymplectic ElementsArXiv2017Prisacariu, VAKähler, OGolodetz, SSapienza, MCavallari, TTorr, PMurray, DWVolumetric models have become a popular representation for 3D scenes in recent years. One breakthrough leading to their popularity was KinectFusion, which focuses on 3D reconstruction using RGB-D sensors. However, monocular SLAM has since also been tackled with very similar approaches. Representing the reconstruction volumetrically as a TSDF leads to most of the simplicity and efficiency that can be achieved with GPU implementations of these systems. However, this representation is memory-intensive and limits applicability to small-scale reconstructions. Several avenues have been explored to overcome this. With the aim of summarizing them and providing for a fast, flexible 3D reconstruction pipeline, we propose a new, unifying framework called InfiniTAM. The idea is that steps like camera tracking, scene representation and integration of new data can easily be replaced and adapted to the user's needs. This report describes the technical implementation details of InfiniTAM v3, the third version of our InfiniTAM system. We have added various new features, as well as making numerous enhancements to the low-level code that significantly improve our camera tracking performance. The new features that we expect to be of most interest are (i) a robust camera tracking module; (ii) an implementation of Glocker et al.'s keyframe-based random ferns camera relocaliser; (iii) a novel approach to globally-consistent TSDF-based reconstruction, based on dividing the scene into rigid submaps and optimising the relative poses between them; and (iv) an implementation of Keller et al.'s surfel-based reconstruction approach. |
| spellingShingle | Prisacariu, VA Kähler, O Golodetz, S Sapienza, M Cavallari, T Torr, P Murray, DW InfiniTAM v3: a framework for large-scale 3D reconstruction with loop closure |
| title | InfiniTAM v3: a framework for large-scale 3D reconstruction with loop closure |
| title_full | InfiniTAM v3: a framework for large-scale 3D reconstruction with loop closure |
| title_fullStr | InfiniTAM v3: a framework for large-scale 3D reconstruction with loop closure |
| title_full_unstemmed | InfiniTAM v3: a framework for large-scale 3D reconstruction with loop closure |
| title_short | InfiniTAM v3: a framework for large-scale 3D reconstruction with loop closure |
| title_sort | infinitam v3 a framework for large scale 3d reconstruction with loop closure |
| work_keys_str_mv | AT prisacariuva infinitamv3aframeworkforlargescale3dreconstructionwithloopclosure AT kahlero infinitamv3aframeworkforlargescale3dreconstructionwithloopclosure AT golodetzs infinitamv3aframeworkforlargescale3dreconstructionwithloopclosure AT sapienzam infinitamv3aframeworkforlargescale3dreconstructionwithloopclosure AT cavallarit infinitamv3aframeworkforlargescale3dreconstructionwithloopclosure AT torrp infinitamv3aframeworkforlargescale3dreconstructionwithloopclosure AT murraydw infinitamv3aframeworkforlargescale3dreconstructionwithloopclosure |