3D Reconstruction and Texture Optimization Using a Sparse Set of RGB-D Cameras
Abstract
We contribute a new integrated system designed for high-quality 3D reconstructions. The system consists of a sparse set of commodity RGB-D cameras, which allows for fast and accurate scan of objects with multi-view inputs. We propose a robust and efficient tile-based streaming pipeline for geometry reconstruction with TSDF fusion which minimizes memory overhead and calculation cost. Our multi-grid warping method for texture optimization can address misalignments of both global structures and small details due to the errors in multi-camera registration, optical distortions and imprecise geometries. In addition, we apply a global color correction method to reduce color inconsistency among RGB images caused by variations of camera settings. Finally, we demonstrate the effectiveness of our proposed system with detailed experiments of multi-view datasets.
Cite
Text
Li et al. "3D Reconstruction and Texture Optimization Using a Sparse Set of RGB-D Cameras." IEEE/CVF Winter Conference on Applications of Computer Vision, 2019. doi:10.1109/WACV.2019.00155Markdown
[Li et al. "3D Reconstruction and Texture Optimization Using a Sparse Set of RGB-D Cameras." IEEE/CVF Winter Conference on Applications of Computer Vision, 2019.](https://mlanthology.org/wacv/2019/li2019wacv-d/) doi:10.1109/WACV.2019.00155BibTeX
@inproceedings{li2019wacv-d,
title = {{3D Reconstruction and Texture Optimization Using a Sparse Set of RGB-D Cameras}},
author = {Li, Wei and Xiao, Xiao and Hahn, James K.},
booktitle = {IEEE/CVF Winter Conference on Applications of Computer Vision},
year = {2019},
pages = {1413-1422},
doi = {10.1109/WACV.2019.00155},
url = {https://mlanthology.org/wacv/2019/li2019wacv-d/}
}