FlashFusion: Real-time Globally Consistent Dense 3D Reconstruction using CPU Computing

 Lei Han, Lu FANG


Fig. 1 Framework of FlashFusion.

Abstract

Aiming at the practical usage of dense 3D reconstruction on portable devices, we propose FlashFusion, a Fast LArge-Scale High-resolution (sub-centimeter level) 3D reconstruction system without the use of GPU computing. It enables globally-consistent localization through a robust yet fast global bundle adjustment scheme, and realizes spatial hashing based volumetric fusion running at 300Hz and rendering at 25Hz via highly efficient valid chunk selection and mesh extraction schemes. Extensive experiments on both real world and synthetic datasets demonstrate that FlashFusion succeeds to enable real- time, globally consistent, high-resolution (5mm), and large-scale dense 3D reconstruction using highly-constrained computation, i.e., the CPU computing on portable device.

Fig. 2 3D model reconstructed online.

Fig. 3 Complexity analysis of FlashFusion.


On-line 3D Reconstruction Demo.


Source Code

Video


Citation:

FlashFusion: Real-time Globally Consistent Dense 3D Reconstruction using CPU computing

@article{han2018flashfusion,   
  title={FlashFusion: Real-time Globally Consistent Dense 3D Reconstruction using CPU computing}, 
  author={Lei, Han and Fang, Lu},   
  organization={Robotics: Science and Systems},   
  year={2018}    }

Supplementary:

Real-time Global Registration for Globally Consistent RGBD SLAM

@article{han2018FastGO,   
  title={Real-time Global Registration for Globally Consistent RGBD SLAM}, 
  author={Lei Han, Lan Xu, Dmytro Bobkov, Eckehard Steinbach, Qionghai Dai and Lu Fang},
  year={2018}    }