基于点线特征融合的低纹理单目视觉同时定位与地图构建算法

doi:10.11772/j.issn.1001-9081.2021050749

《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (7): 2170-2176.DOI: 10.11772/j.issn.1001-9081.2021050749

• 多媒体计算与计算机仿真 • 上一篇

基于点线特征融合的低纹理单目视觉同时定位与地图构建算法

潘高峰, 樊渊(), 汝玉, 郭予超

安徽大学电气工程与自动化学院，合肥 230601

收稿日期:2021-05-11 修回日期:2022-01-10 接受日期:2022-02-18 发布日期:2022-04-06 出版日期:2022-07-10
通讯作者: 樊渊
作者简介:潘高峰（1995—），男，河南许昌人，硕士研究生，主要研究方向：视觉同时定位与地图构建
汝玉（1999—），女，安徽合肥人，主要研究方向：机器人导航与控制
郭予超（1997—），男，河南新郑人，博士研究生，主要研究方向：视觉同时定位与地图构建、机器人控制。
基金资助:
国家重点研发计划项目(2018YFB1305804);国家自然科学基金资助项目(61973002);安徽省自然科学基金资助项目(2008085J32);教育部产学合作协同育人项目(201802259001)

Low-texture monocular visual simultaneous localization and mapping algorithm based on point-line feature fusion

Gaofeng PAN, Yuan FAN(), Yu RU, Yuchao GUO

School of Electrical Engineering and Automation，Anhui University，Hefei Anhui 230601，China

Received:2021-05-11 Revised:2022-01-10 Accepted:2022-02-18 Online:2022-04-06 Published:2022-07-10
Contact: Yuan FAN
About author:PAN Gaofeng， born in 1995， M. S. candidate. His research interests include visual simultaneous localization and mapping.
RU Yu， born in 1999. Her research interests include robot navigation and control.
GUO Yuchao， born in 1997， Ph. D. candidate. His research interests include visual simultaneous localization and mapping， robot control.
Supported by:
National Key Research and Development Program of China(2018YFB1305804);National Natural Science Foundation of China(61973002);Natural Science Foundation of Anhui Province(2008085J32);University-Industry Collaborative Education Program of Ministry of Education(201802259001)

摘要/Abstract

摘要：

当图像因相机快速运动造成模糊或者处在低纹理场景时，仅使用点特征的同步定位与地图构建（SLAM）算法难以跟踪提取足够多的特征点，导致定位精度和匹配鲁棒性较差。而如果造成误匹配，甚至系统都无法工作。针对上述问题，提出了一种基于点线特征融合的低纹理单目SLAM算法。首先，加入了线特征来加强系统稳定性，并解决了点特征算法在低纹理场景中提取不足的问题；然后，对点、线特征提取数量的选择引入了加权的思想，根据场景的丰富程度，对点线特征的权重进行了合理分配。所提算法是在低纹理场景下运行的，因而设置以线特征为主、点特征为辅。在TUM室内数据集上的实验结果表明，与现有的点线特征算法相比，所提算法有效地提高了线特征的匹配精度，使得轨迹误差减小了大约9个百分点，也使得特征提取时间减少了30个百分点，使加入的线特征在低纹理场景中发挥出积极有效的作用，提高了数据整体的准确度和可信度。

关键词: 单目视觉, 点线融合, 线匹配, 低纹理场景, 特征加权

Abstract:

When the image is blurred due to rapid camera movement or in low-texture scenes， the Simultaneous Localization And Mapping （SLAM） algorithm using only point features is difficult to track and extract enough feature points， resulting in poor positioning accuracy and matching robustness. If it causes false matching， even the system cannot work. To solve the problem， a low-texture monocular SLAM algorithm based on point-line feature fusion was proposed. Firstly， the line features were added to enhance the system stability， and the problem of insufficient extraction of point feature algorithm in low texture scenes was solved. Then， the idea of weighting was introduced for the extraction number selection of point and line features， and the weight of point and line features were allocated reasonably according to the richness of the scene. The proposed algorithm ran in low-texture scenes， so the line features were set as the main features and the point features were set as the auxiliary features. Experimental results on the TUM indoor dataset show that compared with the existing point-line feature algorithms， the proposed algorithm can effectively improve the matching precision of the line features， has the trajectory error reduced by about 9 percentage points， and has the feature extraction time reduced by 30 percentage points. As the result， the added line features play a positive and effective role in low-texture scenes， and improve the overall accuracy and reliability of the data.

Key words: monocular vision, point-line fusion, line matching, low-texture scene, feature weighting

中图分类号:

TP242.6

潘高峰, 樊渊, 汝玉, 郭予超. 基于点线特征融合的低纹理单目视觉同时定位与地图构建算法[J]. 计算机应用, 2022, 42(7): 2170-2176.

Gaofeng PAN, Yuan FAN, Yu RU, Yuchao GUO. Low-texture monocular visual simultaneous localization and mapping algorithm based on point-line feature fusion[J]. Journal of Computer Applications, 2022, 42(7): 2170-2176.

图/表 13

图1 TUM数据集里的纹理场景和低纹理场景

Fig. 1 Texture scene and low-texture scene in TUM dataset

图2 空间中线投影与匹配线段的关系

Fig. 2 Relationship between spatial line projection and matched line segments

图3 基于点线特征融合的单目SLAM算法流程

Fig. 3 Flow of monocular SLAM algorithm based on point-line feature fusion

图4 FAST角点提取

Fig. 4 FAST corner point extraction

图5 低纹理场景下的LSD线特征提取

Fig. 5 LSD line feature extraction in low-texture scene

图6 三角化获得地图点深度

Fig. 6 Triangularization to obtain depths of map points

图7 点匹配的效果

Fig. 7 Point matching effects

图8 低纹理场景中的点和线提取

Fig. 8 Point and line extraction in low-texture scene

图9 点匹配和线匹配的效果

Fig. 9 Effects of point matching and line matching in low-texture scene

图10 TUM数据集下的点线稀疏地图

Fig. 10 Point-line sparse map in TUM dataset

图11 相机在TUM低纹理序列中的运动轨迹

Fig. 11 Camera trajectories in TUM low-texture sequences

表1 本文算法与其他算法的绝对轨迹误差对比 (cm)

Tab. 1 Absolute trajectory error comparison of the proposed algorithm and other algorithms

TUM数据集	PL-SLAM	LSD-SLAM	ORB-SLAM2	本文算法
fr3-nostructure-notexture-near-withloop	1.56	1.67	—	1.42
fr3-structure-notexture-near-withloop	0.93	0.87	0.99	0.81
fr3-structure-notexture-far	1.06	1.44	1.55	1.12
fr3-structure-notexture-far-validation	1.34	1.38	1.42	1.23

表2 各个操作的运行时间 (ms)

Tab. 2 Running time of each operation

操作	PL-SLAM	本文算法
特征提取	2.911	1.985
位姿估计	1.000	1.000
点线特征剔除	1.180	1.178
局部BA	1.842	1.769
关键帧剔除	4.441	4.332

参考文献 21

1	DAVISON A J， REID I D， MOLTON N D. MonoSLAM： real-time single camera SLAM［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2007， 29（6）： 1052-1067. 10.1109/tpami.2007.1049
2	张国良，姚二亮. 移动机器人的SLAM与VSLAM方法［M］. 西安：西安交通大学出版社， 2018： 10-15.
	ZHANG G L， YAO E L. SLAM and VSLAM Methods for Mobile Robots ［M］. Xi’an： Xi’an Jiaotong University Press， 2018： 10-15.
3	KLEIN G， MURRAY D. Parallel tracking and mapping for small AR workspaces［C］// Proceedings of the 6th IEEE and ACM International Symposium on Mixed and Augmented Reality. Piscataway： IEEE， 2007： 225-234. 10.1109/ismar.2007.4538852
4	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. 10.1109/tro.2017.2705103
5	ZHANG L L， KOCH R. Hand-held monocular SLAM based on line segments［C］// Proceedings of the 2011 Irish Machine Vision and Image Processing Conference. Piscataway： IEEE， 2011： 7-14. 10.1109/imvip.2011.11
6	ENGEL J， KOLTUN V， CREMERS D. Direct sparse odometry［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2018， 40（3）： 611-625. 10.1109/tpami.2017.2658577
7	ENGEL J， SCHÖPS T， CREMERS D. LSD-SLAM： large-scale direct monocular SLAM［C］// Proceedings of the 2014 European Conference on Computer Vision， LNCS 8690. Cham： Springer， 2014： 834-849.
8	NEWCOMBE R A， LOVEGROVE S J， DAVISON A J. DTAM： dense tracking and mapping in real-time［C］// Proceedings of the 2011 International Conference on Computer Vision. Piscataway： IEEE， 2011： 2320-2327. 10.1109/iccv.2011.6126513
9	PIZZOLI M， FORSTER C， SCARAMUZZA D. REMODE： probabilistic， monocular dense reconstruction in real time［C］// Proceedings of the 2014 IEEE International Conference on Robotics and Automation. Piscataway： IEEE， 2014： 2609-2616. 10.1109/icra.2014.6907233
10	PUMAROLA A， VAKHITOV A， AGUDO A， et al. PL-SLAM： real-time monocular visual SLAM with points and lines［C］// Proceedings of the 2017 IEEE International Conference on Robotics and Automation. Piscataway： IEEE， 2017： 4503-4508. 10.1109/icra.2017.7989522
11	ZUO X X， XIE X J， LIU Y， et al. Robust visual SLAM with point and line features［C］// Proceedings of the 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway： IEEE， 2017： 1775-1782. 10.1109/iros.2017.8205991
12	GOMEZ-OJEDA R， GONZALEZ-JIMENEZ J. Robust stereo visual odometry through a probabilistic combination of points and line segments［C］// Proceedings of the 2016 IEEE International Conference on Robotics and Automation. Piscataway： IEEE， 2016： 2521-2526. 10.1109/icra.2016.7487406
13	ZHANG G X， SUH I H. SoF-SLAM： segments-on-floor-based monocular SLAM［C］// Proceedings of the 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway： IEEE， 2010： 2083-2088.
14	ZHANG G X， SUH I H. Building a partial 3D line-based map using a monocular SLAM［C］// Proceedings of the 2011 IEEE International Conference on Robotics and Automation. Piscataway： IEEE， 2011：1497-1502. 10.1109/icra.2011.5979665
15	R G von GIOI， JAKUBOWICZ J， MOREL J M， et al. LSD： a fast line segment detector with a false detection control［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2010， 32（4）： 722-732. 10.1109/tpami.2008.300
16	ZHANG L L， KOCH R. An efficient and robust line segment matching approach based on LBD descriptor and pairwise geometric consistency［J］. Journal of Visual Communication and Image Representation， 2013， 24（7）：794-805. 10.1016/j.jvcir.2013.05.006
17	高翔，张涛，刘毅，等. 视觉SLAM十四讲：从理论到实践［M］. 北京：电子工业出版社， 2017： 155-157.
	GAO X， ZHANG T， LIU Y， et al. Vision SLAM 14 Lectures： From Theory to Practice［M］. Beijing： Publishing House of Electronics Industry， 2017： 155-157.
18	BARTOLI A， STURM P. The 3D line motion matrix and alignment of line reconstructions［J］. International Journal of Computer Vision， 2004， 57（3）：159-178. 10.1023/b:visi.0000013092.07433.82
19	ROSTEN E， DRUMMOND T. Machine learning for high-speed corner detection［C］// Proceedings of the 2006 European Conference on Computer Vision， LNCS 3951. Berlin： Springer， 2006： 430-443.
20	HARTLEY R， ZISSERMAN A. Multiple View Geometry in Computer Vision［M］. 2nd ed. Cambridge： Cambridge University Press， 2003：33-36. 10.1017/cbo9780511811685
21	席志红，王洪旭，韩双全. 基于ORB-SLAM2系统的快速误匹配剔除算法与地图构建［J］. 计算机应用， 2020， 40（11）： 3289-3294.
	XI Z H， WANG H X， HAN S Q. Fast mismatch elimination algorithm and map-building based on ORB-SLAM2 system［J］. Journal of Computer Applications， 2020， 40（11）： 3289-3294.

[1]	邱云志, 汪廷华, 戴小路. 双重特征加权模糊支持向量机[J]. 《计算机应用》唯一官方网站, 2022, 42(3): 683-687.
[2]	袁蒙恩, 陈立家, 冯子凯. 基于单目视觉的多种群粒子群机械臂路径规划算法[J]. 计算机应用, 2020, 40(10): 2863-2871.
[3]	张午阳, 章伟, 宋芳, 龙林. 基于深度学习的四旋翼无人机单目视觉避障方法[J]. 计算机应用, 2019, 39(4): 1001-1005.
[4]	朱杰, 张俊三, 吴树芳, 董宇坤, 吕琳. 基于多中心卷积特征加权的图像检索方法[J]. 计算机应用, 2018, 38(10): 2778-2781.
[5]	罗余洋, 徐为民, 张梦杰, 刘玉强. 采用单目视觉的桥吊负载空间定位方法[J]. 计算机应用, 2016, 36(4): 1156-1162.
[6]	邱云飞, 刘世兴, 林明明, 邵良杉. 基于分布和逆文本类别指数的特征迁移加权算法[J]. 计算机应用, 2015, 35(6): 1643-1648.
[7]	支晓斌, 许朝晖. 鲁棒的特征权重自调节软子空间聚类算法[J]. 计算机应用, 2015, 35(3): 770-774.
[8]	何少佳刘子扬史剑清. 基于单目视觉的室内机器人障碍检测方案[J]. 计算机应用, 2012, 32(09): 2556-2559.
[9]	刘磊陈兴蜀尹学渊段意吕昭. 基于特征加权朴素贝叶斯算法的网络用户识别[J]. 计算机应用, 2011, 31(12): 3268-3270.
[10]	张翔邓赵红王士同. 具有更好适应性的间距最大化特征加权[J]. 计算机应用, 2010, 30(9): 2275-2278.
[11]	孔攀邓辉文江欢黄艳艳. 改进的基于核函数的模糊聚类算法[J]. 计算机应用, 2008, 28(9): 2338-2340.
[12]	刘志强温华. 基于单目视觉的车辆碰撞预警系统[J]. 计算机应用, 2007, 27(8): 2056-2058.

基于点线特征融合的低纹理单目视觉同时定位与地图构建算法

Low-texture monocular visual simultaneous localization and mapping algorithm based on point-line feature fusion

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 13

参考文献 21

相关文章 12

编辑推荐

Metrics