Three-dimensional reconstruction of large-scale scene based on depth camera

Dong-sheng LIU; Jian-lin CHEN; Dian FEI; Zhi-jiang ZHANG

doi:10.3788/OPE.20202801.0234

您当前的位置：

首页 >

文章列表页 >

Three-dimensional reconstruction of large-scale scene based on depth camera

Information Sciences | 更新时间：2020-08-13

- Three-dimensional reconstruction of large-scale scene based on depth camera
- Optics and Precision Engineering Vol. 28, Issue 1, Pages: 234-243(2020)
- 作者机构：
  
  上海大学特种光纤与光接入网重点实验室，上海 200444
- 作者简介：
- 基金信息：
- DOI：10.3788/OPE.20202801.0234
  CLC： TN249
- Received：01 July 2019，
  
  Accepted：27 August 2019，
  
  Published：25 January 2020
- 稿件说明：
移动端阅览
Dong-sheng LIU, Jian-lin CHEN, Dian FEI, et al. Three-dimensional reconstruction of large-scale scene based on depth camera[J]. Optics and precision engineering, 2020, 28(1): 234-243.
DOI：

Dong-sheng LIU, Jian-lin CHEN, Dian FEI, et al. Three-dimensional reconstruction of large-scale scene based on depth camera[J]. Optics and precision engineering, 2020, 28(1): 234-243. DOI： 10.3788/OPE.20202801.0234.

摘要

针对大场景三维重建中，由位姿估计的累积误差而导致的相机漂移和重建模型质量低的问题，提出了减少累积误差的方法。首先，基于由最新K对深度和彩色图像融合的模型，最小化输入RGB-D图像的几何误差和亮度误差来跟踪相机。然后，若相机位置与当前子网格的中心点距离大于给定阈值时，则将子网格平移体素单元整数倍的距离，基于新建的子网格继续跟踪相机并重建局部场景模型。最后，在子网格间以迭代步长式的方法寻找对应表面点，以对应点间的欧氏距离与亮度误差为约束，优化全局相机轨迹。基于数据集的实验结果表明

相机位姿估计精度比主流方法提升14.1%，全局轨迹优化精度提升8%。对于自采数据，本文设计的系统可减少位姿估计中的累积误差、重建高质量的场景模型。

Abstract

Aiming at the problem of camera drift and low-quality reconstruction caused by accumulated errors in pose estimation for large-scale reconstruction

we proposed a method to reduce accumulated errors. First

based on the model fused by the latest

depth and color images

the geometric and photometric error of the input RGB-D image were minimized to track the camera. Then

if the distance between the camera position and the subvolume center was more than a given threshold

the subvolume was shifted by multiples of voxel size

the camera was continuously tracked and the local scene model was reconstructed based on the newly created subvolume. Finally

the corresponding surface points were searched by the iterative step method between subvolumes

and the global camera trajectory was optimized by Euclidean distance and photometric error between the correspondence. The experimental results based on the dataset show that the camera pose estimation accuracy is improved by 14.1% than the current method

and the global trajectory optimization accuracy is improved by 8%. The system designed in this paper can also reduce the accumulated errors in pose estimation and reconstruct a high-quality scene model for self-collected data.

关键词

Keywords

references

NEWCOMBE R A, IZADI S, HILLIGES O, et al .. Kinect Fusion: Real-Time Dense Surface Mapping and Tracking[C]. IEEE Computer Society , 2011: 127-136.

SZYMON RUSINKIEWICZ, MARC LEVOY. Efficient variants of the icp algorithm[J]. International Conference on 3D Digital Imaging Modeling , 2001:145-152.

H ROTH, M VONA. Moving Volume Kinect Fusion[C]. Proceedings of British Machine Vision Conference , 2012: 1-11.

WHELAN T, KAESS M, JOHANNSSON H, et al .. Real-time large-scale dense RGB-D SLAM with volumetric fusion[J]. International Journal of Robotics Research , 2014, 34(4-5):598-626.

FIORAIO N, TAYLOR J, FITZGIBBON A, et al .. Large-scale and drift-free surface reconstruction using online subvolume registration[J]. IEEE Computer Vision & Pattern Recognition , 2015:4475-4483.

XINGYIN F, FENG Z, QINGXIAO W, et al .. Real-time large-scale dense mapping with surfels[J]. Sensors , 2018, 18(5):1-19.

ZHOU Q Y, MILLER S, KOLTUN V. Elastic Fragments for Dense Scene Reconstruction[C]. IEEE International Conference on Computer Vision , 2013: 473-480.

CHOI N S, ZHOU Q Y, KOLTUN V. Robust Reconstruction of Indoor Scenes[C]. IEEE Conference on Computer Vision and Pattern Recognition , 2015: 5556-5565.

DAI A, MATTHIAS NIEβNER, MICHAEL ZOLLHÖFER, et al .. BundleFusion: real-time globally consistent 3D reconstruction using on-the-fly surface re-integration[J]. ACM Transactions on Graphics , 2017, 36(4): 3054739.

CURLESS B, LEVOY M. A volumetric method for building complex models from range images[J]. In SIGGRAPH International Conference , 1996:303-312.

LORENSEN W E, CLINE H E. Marching cubes: a high resolution 3D surface construction algorithm[J]. ACM Siggraph Computer Graphics , 1987, 21(4):163-169.

ERIK BYLOW, CARL OLSSON, FREDRIK KAHL. Robust Camera Tracking by Combining Color and Depth Measurements[C]. IEEE International Conference on Pattern Recognition , 2014: 4038-4043.

WHELAN T, JOHANNSSON H, KAESS M, et al .. Robust Real-Time Visual Odometry for Dense RGB-D Mapping[C]. IEEE International Conference on Robotics and Automation , 2013: 5724-5731.

BYLOW E, JVRGEN STURM, KERL C, et al .. Real-time camera tracking and 3d reconstruction using signed distance functions[J]. Robotics: Science and Systems , 2013:35-42.

李劲澎, 姜挺, 龚志辉, 等.稳健李代数旋转平均用于GPS辅助无人机影像三维重建[J].光学精密工程, 2017, 25(6): 1607-1618.

ZHANG J Q, JIANG T, GONG ZH H, et al .. GPS-supported 3D reconstruction of UAV images based on robust Lie-algebraic rotation averaging[J]. Opt. Precision Eng ., 2017, 25(6): 1607-1618. (in Chinese)

薛婷, 阮维鹏, 张少杰.基于激光扫描的密集三维重建与优化[J].光学精密工程, 2017, 25(3): 555-561.

XUE T, RUAN W P, ZHANG SH J. 3D reconstruction and optimization of dense bubbly flow based on laser scanning[J]. Opt. Precision Eng ., 2017, 25(3): 555-561. (in Chinese)

高瞻宇, 顾营迎, 吕耀宇, 等.融合通用形变模型信息的面部三维重建[J].光学精密工程, 2018, 26(1): 184-192.

GAO ZH Y, GU Y Y, LV Y Y, et al .. Face reconstruction fused with generic morphable model[J]. Opt. Precision Eng ., 2018, 26(1): 184-192. (in Chinese)

LIU H, LI C, CHEN G, et al .. Robust keyframe-based dense slam with an RGB-D camera[J]. IEEE Computer Vision and Pattern Recognition , 2015:1-12.

FRANK STEINBRVCKER, KERL C, CREMERS D. Large-Scale Multi-resolution Surface Reconstruction from RGB-D Sequences[C]. IEEE International Conference on Computer Vision , 2013: 3264-3271.

NIENER M, ZOLLHFER M, IZADI S, et al .. Real-time 3D reconstruction at scale using voxel hashing[J]. ACM Transactions on Graphics , 2013, 32(6):1-11.

Views

345

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Adaptive octree 3D image reconstruction based on plane patch

LightDiffu DCE： low light image enhancement based on light intensity diffusion

Multi-level feature fusion for camera pose regression

Real-time super-resolution for infrared dynamic object video based on airborne platform

A Transformer-based visual tracker via knowledge distillation

Related Author

Cheng YAO

Caiwen MA

YAN Guanghui

WU Baijing

MA Long

SI Junwen

ZHOU Ziwei

ZHU Deyan

Related Institution

Xi'an Institute of Optics and Precision Mechanics， Chinese Academy of Sciences

University of Chinese Academy of Sciences

School of Electronics & Information Engineering， Lanzhou Jiaotong University

College of Computer and Software Engineeringng，University of Science and Technology Liaoning

College of Astronautics， Nanjing University of Aeronautics and Astronautics

AI问答

Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China Postal code：130033
Tel：0431-86176855 Email：gxjmgc@ciomp.ac.cn
Technical support is provided by Beijing Founder electronics co., LTD 吉ICP备11002662号-17 京公网安备11010802024621
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰