Probabilistic Volumetric Fusion for Dense Monocular SLAM

被引:8
|
作者
Rosinol, Antoni [1 ]
Leonard, John J. [1 ]
Carlone, Luca [1 ]
机构
[1] MIT, Cambridge, MA 02139 USA
来源
2023 IEEE/CVF WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV) | 2023年
关键词
D O I
10.1109/WACV56688.2023.00311
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We present a novel method to reconstruct 3D scenes from images by leveraging deep dense monocular SLAM and fast uncertainty propagation. The proposed approach is able to 3D reconstruct scenes densely, accurately, and in real-time while being robust to extremely noisy depth estimates coming from dense monocular SLAM. Differently from previous approaches, that either use ad-hoc depth filters, or that estimate the depth uncertainty from RGB-D cameras' sensor models, our probabilistic depth uncertainty derives directly from the information matrix of the underlying bundle adjustment problem in SLAM. We show that the resulting depth uncertainty provides an excellent signal to weight the depth-maps for volumetric fusion. Without our depth uncertainty, the resulting mesh is noisy and with artifacts, while our approach generates an accurate 3D mesh with significantly fewer artifacts. We provide results on the challenging Euroc dataset, and show that our approach achieves 92% better accuracy than directly fusing depths from monocular SLAM, and up to 90% improvements compared to the best competing approach.
引用
收藏
页码:3096 / 3104
页数:9
相关论文
共 50 条
  • [1] DeepFactors: Real-Time Probabilistic Dense Monocular SLAM
    Czarnowski, Jan
    Laidlow, Tristan
    Clark, Ronald
    Davison, Andrew J.
    IEEE ROBOTICS AND AUTOMATION LETTERS, 2020, 5 (02) : 721 - 728
  • [2] DRM-SLAM: Towards dense reconstruction of monocular SLAM with scene depth fusion
    Ye, Xinchen
    Ji, Xiang
    Sun, Baoli
    Chen, Shenglun
    Wang, Zhihui
    Li, Haojie
    NEUROCOMPUTING, 2020, 396 (396) : 76 - 91
  • [3] Polarimetric Dense Monocular SLAM
    Yang, Luwei
    Tan, Feitong
    Li, Ao
    Cui, Zhaopeng
    Furukawa, Yasutaka
    Tan, Ping
    2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2018, : 3857 - 3866
  • [4] Dense mapping for monocular-SLAM
    Aguilar-Gonzalez, Abiel
    Arias-Estrada, Miguel
    2016 INTERNATIONAL CONFERENCE ON INDOOR POSITIONING AND INDOOR NAVIGATION (IPIN), 2016,
  • [5] TT-SLAM Dense Monocular SLAM for Planar Environments
    Wang, Xi
    Christie, Marc
    Marchand, Eric
    2021 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA 2021), 2021, : 11690 - 11696
  • [6] SlamCraft: Dense Planar RGB Monocular SLAM
    Rambach, Jason
    Lesur, Paul
    Pagani, Alain
    Stricker, Didier
    PROCEEDINGS OF MVA 2019 16TH INTERNATIONAL CONFERENCE ON MACHINE VISION APPLICATIONS (MVA), 2019,
  • [7] Probabilistic Semi-Dense Mapping from Highly Accurate Feature-Based Monocular SLAM
    Mur-Artal, Raul
    Tardos, Juan D.
    ROBOTICS: SCIENCE AND SYSTEMS XI, 2015,
  • [8] Monocular Dense SLAM with Consistent Deep Depth Prediction
    Yan, Feihu
    Wen, Jiawei
    Li, Zhaoxin
    Zhou, Zhong
    ADVANCES IN COMPUTER GRAPHICS, CGI 2021, 2021, 13002 : 113 - 124
  • [9] Localisation accuracy of semi-dense monocular SLAM
    Schreve, Kristiaan
    du Plessies, Pieter G.
    Raetsch, Matthias
    VIDEOMETRICS, RANGE IMAGING, AND APPLICATIONS XIV, 2017, 10332
  • [10] Probabilistic Depth Map Model for Rotation-Only Camera Motion in Semi-Dense Monocular SLAM
    Zhou, Yao
    Yan, Feihu
    Zhou, Zhong
    2016 INTERNATIONAL CONFERENCE ON VIRTUAL REALITY AND VISUALIZATION (ICVRV 2016), 2016, : 8 - 15