单目视觉惯性定位的IMU辅助跟踪模型

doi:10.13474/j.cnki.11-2246.2018.0350

测绘通报 ›› 2018, Vol. 0 ›› Issue (11): 58-62.doi: 10.13474/j.cnki.11-2246.2018.0350

单目视觉惯性定位的IMU辅助跟踪模型

王帅, 潘树国, 黄砺枭, 曾攀

东南大学仪器科学与工程学院, 江苏南京 210096

收稿日期:2018-02-28 修回日期:2018-06-14 出版日期:2018-11-25 发布日期:2018-11-29
作者简介:王帅(1993-),男,硕士生,主要研究方向为视觉惯性紧耦合定位。E-mail:605096939@qq.com
基金资助:
国家自然科学基金（41574026；41774027）；江苏省重点研发计划（BE2016176）；国家重点研发计划（2016YFB0502101）；江苏省六大人才高峰计划（2015-WLW-002）

IMU-assisted Tracking Model of Monocular Vision Inertial Positioning

WANG Shuai, PAN Shuguo, HUANG Lixiao, ZENG Pan

School of Instrument Science and Engineering, Southeast University, Nanjing 210096, China

Received:2018-02-28 Revised:2018-06-14 Online:2018-11-25 Published:2018-11-29

摘要/Abstract

摘要：

针对剧烈运动下单目视觉惯性定位精度较差的问题，提出了一种面向单目视觉惯性紧耦合定位的IMU辅助跟踪模型，以提升单目视觉惯性紧耦合定位下的稳健性。以IMU辅助跟踪模型取代常规的参考帧和匀速跟踪模型，该模型分为两个阶段，初始化阶段时，在匀速模型设定当前帧初始平移的基础上，由IMU预积分确定当前帧的初始旋转，从而获得当前帧的初始位姿；初始化完成后，在提供初始位姿的基础上，加入IMU预积分的先验速度信息；最后由以上计算的初始状态建立跟踪模型，实现精确定位。采用公开的室内SLAM数据集进行验证，结果表明，该IMU辅助跟踪模型可有效提高系统的稳健性，同时定位精度控制在0.1 m左右，其精度相比于传统的跟踪模型约提高20%。

关键词: 单目视觉惯性, IMU辅助跟踪模型, 初始状态, 稳健性

Abstract:

In order to solve the outstanding problem of poor monocular vision inertial positioning accuracy under violent motion,an IMU-assisted tracking model for monocular vision inertial tightly-coupled positioning is proposed to improve the robustness under monocular vision inertial tightly-coupled positioning.The IMU-assisted tracking model is used to replace the conventional reference frame and constant speed tracking model.The model is divided into two phases.In the initialization phase,on the basis of setting the initial translation of the current frame in the constant model,the IMU pre-integration is used to determine the initial rotation of the current frame so as to obtain the initial pose of the current frame.After the initialization is completed,the prior velocity information of the IMU pre-integration is added based on the initial pose provided.Finally,the tracking model is established from the initial state calculated above to achieve accurate positioning.Validation uses open indoor SLAM data sets.The results show that the proposed IMU-assisted tracking model can effectively improve the robustness of the system,while the positioning accuracy is controlled in about 0.1 m,the accuracy is about 20% higher than that of the traditional tracking model.

Key words: monocular visual inertia, IMU-assisted tracking model, initial state, robustness

中图分类号:

王帅, 潘树国, 黄砺枭, 曾攀. 单目视觉惯性定位的IMU辅助跟踪模型[J]. 测绘通报, 2018, 0(11): 58-62.

WANG Shuai, PAN Shuguo, HUANG Lixiao, ZENG Pan. IMU-assisted Tracking Model of Monocular Vision Inertial Positioning[J]. 测绘通报, 2018, 0(11): 58-62.

参考文献

[1] 胡劲草.室内自主式移动机器人定位方法研究[J].传感器世界,2006,20(11):14-18.
[2] 李群明,熊蓉,褚健.室内自主移动机器人定位方法研究综述[J].机器人,2003,25(6):560-567.
[3] PAN S G,MENG X,WANG S L,et al.Ambiguity Resolution with Double Troposphere Parameter Restriction for Long Range Reference Stations in NRTK System[J].Empire Survey Review,2015,47(345):429-437.
[4] PAN S,MENG X,WANG G,et al.A New Approach for Optimising GNSS Positioning Performance in Harsh Observation Environments[J].Journal of Navigation,2014,67(6):1029-1048.
[5] 姜卫平.卫星导航定位基准站网的发展现状、机遇与挑战[J].测绘学报,2017,46(10):1379-1388.
[6] 王勇,陈国良,李晓园,等.一种相机标定辅助的单目视觉室内定位方法[J].测绘通报,2018(2):35-40.
[7] JULIER S J,UHLMANN J K.A New Extension of the Kalman Filter to Nonlinear Systems[C]//Proceedings of SPIE.[S.l.]:SPIE,1997.
[8] TRIGGS B,MCLAUCHLAN P F,HARTLEY R I,et al.Bundle Adjustment-A Modern Synthesis[C]//International Workshop on Vision Algorithms.Berlin:Springer,1999:298-372.
[9] MOURIKIS A I,ROUMELIOTIS S I.A Multi-state Constraint Kalman Filter for Vision-aided Inertial Navigation[C]//Proceedings 2007 IEEE International Conference on Robotics and Automation. Roma:IEEE,2007:3565-3572.
[10] MUR-ARTAL R,TARDÓS J D.Orb-slam2:An Open-source Slam System for Monocular,Stereo,and Rgb-d cameras[J].IEEE Transactions on Robotics,2017,33(5):1255-1262.
[11] MUR-ARTAL R,TARDÓS J D.Visual-inertial Monocular SLAM with Map Reuse[J].IEEE Robotics and Automation Letters,2017,2(2):796-803.
[12] LEE D,KIM H,MYUNG H.2D Image Feature-Based Real-Time RGB-D 3D SLAM[J].Springer Berlin Heidelberg,2013,208:485-492.
[13] QIN T,LI P,SHEN S.Vins-mono:A Robust and Versatile Monocular Visual-inertial State Estimator[J].ArXiv Preprint,2017,8:210-235.
[14] FORSTER C,CARLONE L,DELLAERT F,et al.On-manifold Preintegration for Real-time Visual-inertial Odometry[J].IEEE Transactions on Robotics,2017,33(1):1-21.
[15] LEUTENEGGER S,LYNEN S,BOSSE M,et al.Keyframe-based Visual-inertial Odometry Using Nonlinear Optimization[J].The International Journal of Robotics Research,2015,34(3):314-334.

单目视觉惯性定位的IMU辅助跟踪模型

IMU-assisted Tracking Model of Monocular Vision Inertial Positioning

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 3

编辑推荐

Metrics

本文评价

[1]	普东东, 丁海勇. 融合超体素与成对链接聚类的LiDAR点云分割算法[J]. 测绘通报, 2022, 0(3): 65-69.
[2]	曾攀, 潘树国, 黄砺枭, 王帅, 赵涛. 基于加权预积分和快速初始化的惯性辅助单目前端模型[J]. 测绘通报, 2019, 0(8): 8-13,19.
[3]	陈向阳, 向云飞. 利用LMedS算法提取规则建筑物顶部面片[J]. 测绘通报, 2018, 0(4): 121-124.