Stereo visual SLAM algorithm for fusing point-line-plane features in low texture environments

doi:10.3785/j.issn.1008-973X.2026.02.010

Journal of ZheJiang University (Engineering Science)

2026, Vol. 60

Issue (2): 322-331 DOI: 10.3785/j.issn.1008-973X.2026.02.010

Stereo visual SLAM algorithm for fusing point-line-plane features in low texture environments

Ze WANG1(

),Lei RAO1,*(

),Guangyu FAN1,Niansheng CHEN1,Songlin CHENG1,Dingyu YANG2,Chuqiao JIANG1

1. School of Electronic Information Engineering, Shanghai Dianji University, Shanghai 201306, China
2. State Key Laboratory of Blockchain and Data Security, Zhejiang University, Hangzhou 310058, China

Download:

HTML

PDF(2028KB) HTML
Export: BibTeX | EndNote (RIS)

Abstract

To address the issues of low positioning accuracy and significant trajectory drift errors in ORB-SLAM2 based on point features under low-texture scenes, a stereo visual SLAM algorithm integrating point, line, and plane features was proposed. An improved EDLines line feature extraction algorithm was designed and introduced into ORB-SLAM2. By employing short-line suppression and similar-line merging strategies, computational time was reduced while enhancing the quality of line feature extraction. A plane features extraction method based on intersecting lines was proposed. Geometric constraints derived from extracted plane features were used to optimize pose estimation, and reprojection errors were reduced. A joint optimization method for point, line, and plane features was introduced, integrating geometric relationships across multiple features to mitigate cumulative errors from single-feature reliance. The proposed algorithm’s effectiveness was validated on KITTI, EuRoC, and UMA-VI datasets. Experimental results demonstrate that compared to ORB-SLAM2, point-line feature-based SLAM, and point-plane feature-based SLAM algorithms, the proposed method achieves superior positioning accuracy and robustness.

Key words： low-texture environment visual SLAM line features plane features joint optimization

Received: 15 February 2025 Published: 03 February 2026

CLC:

TP 242

Fund: 国家自然科学基金资助项目（61702320）；上海市晨光计划（15CG62）.

Corresponding Authors: Lei RAO E-mail: 949795713@qq.com;raol@sdju.edu.cn

	Service
	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	Ze WANG
	Lei RAO
	Guangyu FAN
	Niansheng CHEN
	Songlin CHENG
	Dingyu YANG
	Chuqiao JIANG

Cite this article:

Ze WANG,Lei RAO,Guangyu FAN,Niansheng CHEN,Songlin CHENG,Dingyu YANG,Chuqiao JIANG. Stereo visual SLAM algorithm for fusing point-line-plane features in low texture environments. Journal of ZheJiang University (Engineering Science), 2026, 60(2): 322-331.

URL:

https://www.zjujournals.com/eng/10.3785/j.issn.1008-973X.2026.02.010 OR https://www.zjujournals.com/eng/Y2026/V60/I2/322

低纹理环境下融合点线面特征的双目视觉SLAM算法

针对机器人在低纹理场景下基于点特征的ORB-SLAM2存在定位精度低、轨迹漂移误差较大的问题，提出融合点线面特征的双目视觉SLAM算法. 在ORB-SLAM2中设计并引入改进的EDLines线特征提取算法，通过短线抑制和相似直线合并策略，降低计算时间并提高线特征提取的质量. 提出基于相交直线的平面特征提取方法，基于所提取面特征的几何约束优化位姿估计，减少重投影误差. 提出点线面特征的联合优化方法，融合多种特征的几何关系，减少由单一特征带来的误差累积. 在KITTI、EuRoC和UMA-VI数据集下测试所提算法的有效性. 实验结果表明，相较于ORB-SLAM2、点线特征SLAM以及点面特征SLAM算法，所提算法在定位精度与鲁棒性方面更优.

关键词： 低纹理环境, 视觉SLAM, 线特征, 面特征, 联合优化

Fig.1 Framework of stereo visual SLAM algorithm for fusing point-line-plane features

Fig.2 Merging of similar line segments

Fig.3 Performance comparison of line feature extraction between two algorithms

Fig.4 Matching of line segments and endpoints

Fig.5 Plane features extraction

Tab.1 Number of point, line and plane features across different scenes

Fig.6 Features extraction

Tab.2 Comparison of translation root-mean-square error for different visual SLAM algorithms on KITTI dataset

Fig.7 Pose-trajectory comparison of different visual SLAM algorithms on two KITTI-dataset sequences

Tab.3 Comparison of translation root-mean-square error for different visual SLAM algorithms on EuRoC dataset

Fig.8 Pose-trajectory comparison of different visual SLAM algorithms on sequence V2_01_easy

Fig.9 Feature map of sequence V1_02_medium

Tab.4 Comparison of translation root-mean-square error for different visual SLAM algorithms on UMA-VI dataset

Fig.10 Pose-trajectory comparison of different visual SLAM algorithms on sequence corridor-eng


[1]	CADENA C, CARLONE L, CARRILO H, et al Past, present, and future of simultaneous localization and mapping: toward the robust-perception age[J]. IEEE Transactions on Robotics, 2016, 32 (6): 1309- 1332 doi: 10.1109/TRO.2016.2624754

[2]	TAKETOMI T, UCHIYAMA H, IKEDA S Visual SLAM algorithms: a survey from 2010 to 2016[J]. IPSJ Transactions on Computer Vision and Applications, 2017, 9: 16 doi: 10.1186/s41074-017-0027-2

[3]	ENGEL J, KOLTUN V, CREMERS D Direct sparse odometry[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018, 40 (3): 611- 625 doi: 10.1109/TPAMI.2017.2658577

[4]	MUR-ARTAL R, MONTIEL J, TARDÓS J ORB-SLAM: a versatile and accurate monocular SLAM system[J]. IEEE Transactions on Robotics, 2015, 31 (5): 1147- 1163 doi: 10.1109/TRO.2015.2463671

[5]	MUR-ARTAL R, TARDÓS J ORB-SLAM2: an open-source SLAM system for monocular, stereo, and RGB-D cameras[J]. IEEE Transactions on Robotics, 2017, 33 (5): 1255- 1262 doi: 10.1109/TRO.2017.2705103

[6]	CAMPOS C, ELVIRA R, RODRÍGUEZ J, et al ORB-SLAM3: an accurate open-source library for visual, visual–inertial, and multimap SLAM[J]. IEEE Transactions on Robotics, 2021, 37 (6): 1874- 1890 doi: 10.1109/TRO.2021.3075644

[7]	WEI H, TANG F, XU Z, et al A point-line VIO system with novel feature hybrids and with novel line predicting-matching[J]. IEEE Robotics and Automation Letters, 2021, 6 (4): 8681- 8688 doi: 10.1109/LRA.2021.3113987

[8]	KAESS M. Simultaneous localization and mapping with infinite planes [C]// Proceedings of the 2015 IEEE International Conference on Robotics and Automation. Seattle: IEEE, 2015: 4605–4611.

[9]	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. Singapore: IEEE, 2017: 4503–4508.

[10]	GIOI R, JAKUBOWICZ J, MOREL J, 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 doi: 10.1109/TPAMI.2008.300

[11]	QIAN K, ZHAO W, LI K, et al Visual SLAM with BoPLW pairs using egocentric stereo camera for wearable-assisted substation inspection[J]. IEEE Sensors Journal, 2020, 20 (3): 1630- 1641 doi: 10.1109/JSEN.2019.2947275

[12]	TREVOR A J B, GEDIKLI S, RUSU R B, et al Efficient organized point cloud segmentation with connected components[J]. Semantic Perception Mapping Exploration, 2013, 10 (6): 251- 257

[13]	MA L, KERL C, STÜCKLER J, et al. CPA-SLAM: consistent plane-model alignment for direct RGB-D SLAM [C]// Proceedings of the 2016 IEEE International Conference on Robotics and Automation. Stockholm: IEEE, 2016: 1285–1291.

[14]	ZHOU L, KOPPEL D, KAESS M Lidar SLAM with plane adjustment for indoor environment[J]. IEEE Robotics and Automation Letters, 2021, 6 (4): 7073- 7080 doi: 10.1109/LRA.2021.3092274

[15]	ZHANG X, WANG W, QI X, et al. Stereo plane SLAM based on intersecting lines [C]// Proceedings of the 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems. Prague: IEEE, 2021: 6566–6572.

[16]	黄宁生, 陈靖, 缪远东基于平面约束的RGB-D SLAM系统[J]. 计算机应用研究, 2020, 37 (8): 2526- 2530 HUANG Ningsheng, CHEN Jing, MIAO Yuandong RGB-D SLAM system based on plane constraint[J]. Application Research of Computers, 2020, 37 (8): 2526- 2530 doi: 10.19734/j.issn.1001-3695.2019.03.0100

[17]	ZHANG X, WANG W, QI X, et al Point-plane SLAM using supposed planes for indoor environments[J]. Sensors, 2019, 19 (17): 3795 doi: 10.3390/s19173795

[18]	WANG H, WEI H, XU Z, et al RSS: robust stereo SLAM with novel extraction and full exploitation of plane features[J]. IEEE Robotics and Automation Letters, 2024, 9 (6): 5158- 5165 doi: 10.1109/LRA.2024.3388854

[19]	YAN J, ZHENG Y, YANG J, et al PLPF-VSLAM: an indoor visual SLAM with adaptive fusion of point-line-plane features[J]. Journal of Field Robotics, 2024, 41 (1): 50- 67 doi: 10.1002/rob.22242

[20]	SHU F, WANG J, PAGANI A, et al. Structure PLP-SLAM: efficient sparse mapping and localization using point, line and plane for monocular, RGB-D and stereo cameras [C]// Proceedings of the 2023 IEEE International Conference on Robotics and Automation. London: IEEE, 2023: 2105–2112.

[21]	AKINLAR C, TOPAL C EDLines: a real-time line segment detector with a false detection control[J]. Pattern Recognition Letters, 2011, 32 (13): 1633- 1642 doi: 10.1016/j.patrec.2011.06.001

[22]	ZHANG 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 doi: 10.1016/j.jvcir.2013.05.006

[23]	BURRI M, NIKOLIC J, GOHL P, et al The EuRoC micro aerial vehicle datasets[J]. International Journal of Robotics Research, 2016, 35 (10): 1157- 1163 doi: 10.1177/0278364915620033

[24]	GEIGER A, LENZ P, STILLER C, et al Vision meets robotics: the KITTI dataset[J]. International Journal of Robotics Research, 2013, 32 (11): 1231- 1237 doi: 10.1177/0278364913491297

[1]	Kai LIN,Xin-wu LIANG,Ji-yuan CAI. Dynamic RGB-D SLAM algorithm based on reprojection depth difference cumulative map and static probability[J]. Journal of ZheJiang University (Engineering Science), 2022, 56(6): 1062-1070.

[2]	Xin MA,Xin-wu LIANG,Ji-yuan CAI. Fast visual SLAM method based on point and line features[J]. Journal of ZheJiang University (Engineering Science), 2021, 55(2): 402-409.

[3]	CHEN Xun, ZHANG Chao-Yang, LUO Hai-Yan. Joint optimization of power control, channel assignment and scheduling in wireless mesh network[J]. Journal of ZheJiang University (Engineering Science), 2009, 43(8): 1406-1411.

Viewed

Full text

Abstract

Cited

Shared

Discussed