ViPR: Visual-Odometry-aided Pose Regression for 6DoF Camera Localization

被引:6
|
作者
Ott, Felix [1 ]
Feigl, Tobias [1 ,2 ]
Loeffler, Christoffer [1 ,2 ]
Mutschler, Christopher [1 ,3 ]
机构
[1] Fraunhofer Inst Integrated Circuits IIS, Nurnberg, Germany
[2] FAU Erlangen Nuremberg, Dept Comp Sci, Erlangen, Germany
[3] Ludwig Maximilians Univ LMU, Dept Stat, Munich, Germany
来源
2020 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS (CVPRW 2020) | 2020年
关键词
AUGMENTED REALITY; EFFICIENT;
D O I
10.1109/CVPRW50498.2020.00029
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Visual Odometry (VO) accumulates a positional drift in long-term robot navigation tasks. Although Convolutional Neural Networks (CNNs) improve VO in various aspects, VO still suffers from moving obstacles, discontinuous observation of features, and poor textures or visual information. While recent approaches estimate a 6DoF pose either directly from (a series of) images or by merging depth maps with optical flow (OF), research that combines absolute pose regression with OF is limited. We propose ViPR, a novel modular architecture for long-term 6DoF VO that leverages temporal information and synergies between absolute pose estimates (from PoseNet-like modules) and relative pose estimates (from FlowNet-based modules) by combining both through recurrent layers. Experiments on known datasets and on our own Industry dataset show that our modular design outperforms state of the art in long-term navigation tasks.
引用
收藏
页码:187 / 198
页数:12
相关论文
共 50 条
  • [1] Normalized Localization for 6-DOF Camera Pose Regression
    Huang, Weiquan
    Bai, Yan
    Wang, Yixin
    Wu, Yutang
    Feng, Ming
    Wang, Yin
    ELEVENTH INTERNATIONAL CONFERENCE ON GRAPHICS AND IMAGE PROCESSING (ICGIP 2019), 2020, 11373
  • [2] Precision potential of photogrammetric 6DOF pose estimation with a single camera
    Luhmann, Thomas
    ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING, 2009, 64 (03) : 275 - 284
  • [3] Summarizing image/surface registration for 6DOF robot/camera pose estimation
    Batlle, Elisabet
    Matabosch, Carles
    Salvi, Joaquim
    PATTERN RECOGNITION AND IMAGE ANALYSIS, PT 2, PROCEEDINGS, 2007, 4478 : 105 - +
  • [4] Benchmarking 6DOF Outdoor Visual Localization in Changing Conditions
    Sattler, Torsten
    Maddern, Will
    Toft, Carl
    Torii, Akihiko
    Hammarstrand, Lars
    Stenborg, Erik
    Safari, Daniel
    Okutomi, Masatoshi
    Pollefeys, Marc
    Sivic, Josef
    Kahl, Fredrik
    Pajdla, Tomas
    2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2018, : 8601 - 8610
  • [5] Visual object tracking by a camera mounted on a 6DOF industrial robot
    Aouf, N
    Rajabi, H
    Rajabi, N
    Alanbari, H
    Perron, C
    2004 IEEE CONFERENCE ON ROBOTICS, AUTOMATION AND MECHATRONICS, VOLS 1 AND 2, 2004, : 213 - 218
  • [6] Plane-Aided Visual-Inertial Odometry for 6-DOF Pose Estimation of a Robotic Navigation Aid
    Zhang, He
    Ye, Cang
    IEEE ACCESS, 2020, 8 : 90042 - 90051
  • [7] "Recent Methods of 6DoF Pose Estimation"
    Akizuki S.
    Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers, 2019, 73 (02): : 210 - 213
  • [8] Single-Camera Multi-View 6DoF pose estimation for robotic grasping
    Yuan, Shuangjie
    Ge, Zhenpeng
    Yang, Lu
    FRONTIERS IN NEUROROBOTICS, 2023, 17
  • [9] Affine Correspondences Between Multi-camera Systems for 6DOF Relative Pose Estimation
    Guan, Banglei
    Zhao, Ji
    COMPUTER VISION - ECCV 2022, PT XXXII, 2022, 13692 : 634 - 650
  • [10] A Benchmark Dataset for 6DoF Object Pose Tracking
    Wu, Po-Chen
    Lee, Yueh-Ying
    Tseng, Hung-Yu
    Ho, Hsuan-I
    Yang, Ming-Hsuan
    Chien, Shao-Yi
    ADJUNCT PROCEEDINGS OF THE 2017 IEEE INTERNATIONAL SYMPOSIUM ON MIXED AND AUGMENTED REALITY (ISMAR-ADJUNCT), 2017, : 186 - 191