Robot dynamic scene visual SLAM algorithm integrating geometric features and semantic information

doi:10.13474/j.cnki.11-2246.2025.0911

Abstract

Abstract: To address the challenges of achieving high-accuracy visual SLAM in dynamic scenes,including low robustness,large localization errors,and feature loss,this paper proposes a visual SLAM algorithm that integrates geometric features and semantic information (Geo-Semantic SLAM).Built upon the ORB-SLAM2 framework,the proposed method deeply integrates semantic segmentation networks with geometric feature extraction techniques,significantly enhancing its adaptability to dynamic environments.Specifically,Geo-Semantic SLAM employs semantic segmentation to eliminate the influence of dynamic objects and introduces a camera pose optimization method based on the semantic information of static objects,effectively compensating for the feature loss caused by dynamic object removal.Experimental validation on the TUM dataset demonstrates that Geo-Semantic SLAM achieves superior localization and mapping accuracy in dynamic environments.Moreover,it consistently outperforms traditional algorithms and semantic SLAM methods across various scenarios.

Key words: geometric features, semantic information, dynamic scenes, robots, SLAM

CLC Number:

P237

HE Tingting, JIANG Xianglong, HE Shengxi. Robot dynamic scene visual SLAM algorithm integrating geometric features and semantic information[J]. Bulletin of Surveying and Mapping, 2025, 0(9): 64-69.

References

[1] CHEN Weifeng, SHANG Guangtao, JI Aihong, et al.An overview on visual SLAM: from tradition to semantic[J].Remote Sensing, 2022, 14(13):3010.
[2] BOKOVOY A, MURAVIEV K.Assessment of map construction in vSLAM[C]//Proceedings of 2021 International Siberian Conference on Control and Communications (SIBCON).Kazan:IEEE, 2021:1-6.
[3] CADENA C, CARLONE L, CARRILLO 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.
[4] MUR-ARTAL R, TARDOS 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.
[5] LIU R, WANG X, CHEN Z, et al.A survey on visual SLAM based on deep learning[J].Journal of System Simulation, 2020, 32(7):1244-1256.
[6] YU Chao, LIU Zuxin, LIU Xinjun, et al.DS-SLAM:a semantic visual SLAM towards dynamic environments[C]//Proceedings of 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).Madrid:IEEE, 2018:1168-1174.
[7] BESCOS B, FACIL J M, CIVERA J, et al.DynaSLAM:tracking, mapping, and inpainting in dynamic scenes[J].IEEE Robotics and Automation Letters, 2018, 3(4):4076-4083.
[8] CHENG Shuhong, SUN Changhe, ZHANG Shijun, et al.SG-SLAM:a real-time RGB-D visual SLAM toward dynamic scenes with semantic and geometric information[J].IEEE Transactions on Instrumentation Measurement, 2023, 72:3228006.
[9] LIU Yubao, MIURA J.RDS-SLAM:real-time dynamic SLAM using semantic segmentation methods[J].IEEE Access, 2021, 9:23772-23785.
[10] ESPARZA D, FLORES G.The STDyn-SLAM:a stereo vision and semantic segmentation approach for VSLAM in dynamic outdoor environments[J].IEEE Access, 2022, 10:18201-18209.
[11] BADRINARAYANAN V, KENDALL A, CIPOLLA R.SegNet:a deep convolutional encoder-decoder architecture for image segmentation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(12):2481-2495.
[12] MERRILL N, GUO Yuliang, ZUO Xingxing, et al.Symmetry and uncertainty-aware object SLAM for 6DoF object pose estimation[C]//Proceedings of 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).New Orleans, LA:IEEE, 2022:14881-14890.
[13] JUNG J H, PARK C G.A framework for visual-inertial object-level simultaneous localization and mapping[C]//Proceedings of 2023 IEEE/ION Position, Location and Navigation Symposium (PLANS).Monterey, CA:IEEE, 2023:1335-1340.
[14] LIAO Ziwei, HU Yutong, ZHANG Jiadong, et al.SO-SLAM:semantic object SLAM with scale proportional and symmetrical texture constraints[J].IEEE Robotics and Automation Letters, 2022, 7(2):4008-4015.
[15] ZINS M, SIMON G, BERGER M O.OA-SLAM:leveraging objects for camera relocalization in visual SLAM[C]//Proceedings of 2022 IEEE International Symposium on Mixed and Augmented Reality (ISMAR).Singapore:IEEE, 2022:720-728.
[16] CHEN Jiahao, LI Xiuzhi.EO-SLAM:evolutionary object slam in perceptual constrained scene[M]//Artificial Intelligence.Cham:Springer Nature Switzerland, 2022:426-438.
[17] JIAO Songming, LI Yan, SHAN Zhengwen.DFS-SLAM:a visual SLAM algorithm for deep fusion of semantic information[J].IEEE Robotics and Automation Letters, 2024, 9(12):11794-11801.
[18] HE Kaiming, GKIOXARI G, DOLLAR P, et al.Mask R-CNN[C]//Proceedings of 2017 IEEE International Conference on Computer Vision (ICCV).Venice:IEEE, 2017:2980-2988.
[19] STURM J, ENGELHARD N, ENDRES F, et al.A benchmark for the evaluation of RGB-D SLAM systems[C]//Proceedings of 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems.Vilamoura-Algarve:IEEE, 2012:573-580.