DM-SLAM: A Feature-Based SLAM System for Rigid Dynamic Scenes
Most Simultaneous Localization and Mapping (SLAM) methods assume that environments are static. Such a strong assumption limits the application of most visual SLAM systems. The dynamic objects will cause many wrong data associations during the SLAM process. To address this problem, a novel visual SLA...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-03-01
|
| Series: | ISPRS International Journal of Geo-Information |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2220-9964/9/4/202 |
| _version_ | 1797626505862316032 |
|---|---|
| author | Junhao Cheng Zhi Wang Hongyan Zhou Li Li Jian Yao |
| author_facet | Junhao Cheng Zhi Wang Hongyan Zhou Li Li Jian Yao |
| author_sort | Junhao Cheng |
| collection | DOAJ |
| description | Most Simultaneous Localization and Mapping (SLAM) methods assume that environments are static. Such a strong assumption limits the application of most visual SLAM systems. The dynamic objects will cause many wrong data associations during the SLAM process. To address this problem, a novel visual SLAM method that follows the pipeline of feature-based methods called DM-SLAM is proposed in this paper. DM-SLAM combines an instance segmentation network with optical flow information to improve the location accuracy in dynamic environments, which supports monocular, stereo, and RGB-D sensors. It consists of four modules: semantic segmentation, ego-motion estimation, dynamic point detection and a feature-based SLAM framework. The semantic segmentation module obtains pixel-wise segmentation results of potentially dynamic objects, and the ego-motion estimation module calculates the initial pose. In the third module, two different strategies are presented to detect dynamic feature points for RGB-D/stereo and monocular cases. In the first case, the feature points with depth information are reprojected to the current frame. The reprojection offset vectors are used to distinguish the dynamic points. In the other case, we utilize the epipolar constraint to accomplish this task. Furthermore, the static feature points left are fed into the fourth module. The experimental results on the public TUM and KITTI datasets demonstrate that DM-SLAM outperforms the standard visual SLAM baselines in terms of accuracy in highly dynamic environments. |
| first_indexed | 2024-03-11T10:11:22Z |
| format | Article |
| id | doaj.art-f2bc5675cc4748de8aca924e841369da |
| institution | Directory Open Access Journal |
| issn | 2220-9964 |
| language | English |
| last_indexed | 2024-03-11T10:11:22Z |
| publishDate | 2020-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | ISPRS International Journal of Geo-Information |
| spelling | doaj.art-f2bc5675cc4748de8aca924e841369da2023-11-16T14:31:09ZengMDPI AGISPRS International Journal of Geo-Information2220-99642020-03-019420210.3390/ijgi9040202DM-SLAM: A Feature-Based SLAM System for Rigid Dynamic ScenesJunhao Cheng0Zhi Wang1Hongyan Zhou2Li Li3Jian Yao4School of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430070, ChinaSchool of Software Engineering, Xi’an Jiaotong University, Xi’an 710049, ChinaSchool of Resource and Environment Sciences, Wuhan University, Wuhan 430070, ChinaSchool of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430070, ChinaSchool of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430070, ChinaMost Simultaneous Localization and Mapping (SLAM) methods assume that environments are static. Such a strong assumption limits the application of most visual SLAM systems. The dynamic objects will cause many wrong data associations during the SLAM process. To address this problem, a novel visual SLAM method that follows the pipeline of feature-based methods called DM-SLAM is proposed in this paper. DM-SLAM combines an instance segmentation network with optical flow information to improve the location accuracy in dynamic environments, which supports monocular, stereo, and RGB-D sensors. It consists of four modules: semantic segmentation, ego-motion estimation, dynamic point detection and a feature-based SLAM framework. The semantic segmentation module obtains pixel-wise segmentation results of potentially dynamic objects, and the ego-motion estimation module calculates the initial pose. In the third module, two different strategies are presented to detect dynamic feature points for RGB-D/stereo and monocular cases. In the first case, the feature points with depth information are reprojected to the current frame. The reprojection offset vectors are used to distinguish the dynamic points. In the other case, we utilize the epipolar constraint to accomplish this task. Furthermore, the static feature points left are fed into the fourth module. The experimental results on the public TUM and KITTI datasets demonstrate that DM-SLAM outperforms the standard visual SLAM baselines in terms of accuracy in highly dynamic environments.https://www.mdpi.com/2220-9964/9/4/202visual SLAMdeep learningdynamic scenesMask R-CNNoptical flowORB-SLAM2 |
| spellingShingle | Junhao Cheng Zhi Wang Hongyan Zhou Li Li Jian Yao DM-SLAM: A Feature-Based SLAM System for Rigid Dynamic Scenes ISPRS International Journal of Geo-Information visual SLAM deep learning dynamic scenes Mask R-CNN optical flow ORB-SLAM2 |
| title | DM-SLAM: A Feature-Based SLAM System for Rigid Dynamic Scenes |
| title_full | DM-SLAM: A Feature-Based SLAM System for Rigid Dynamic Scenes |
| title_fullStr | DM-SLAM: A Feature-Based SLAM System for Rigid Dynamic Scenes |
| title_full_unstemmed | DM-SLAM: A Feature-Based SLAM System for Rigid Dynamic Scenes |
| title_short | DM-SLAM: A Feature-Based SLAM System for Rigid Dynamic Scenes |
| title_sort | dm slam a feature based slam system for rigid dynamic scenes |
| topic | visual SLAM deep learning dynamic scenes Mask R-CNN optical flow ORB-SLAM2 |
| url | https://www.mdpi.com/2220-9964/9/4/202 |
| work_keys_str_mv | AT junhaocheng dmslamafeaturebasedslamsystemforrigiddynamicscenes AT zhiwang dmslamafeaturebasedslamsystemforrigiddynamicscenes AT hongyanzhou dmslamafeaturebasedslamsystemforrigiddynamicscenes AT lili dmslamafeaturebasedslamsystemforrigiddynamicscenes AT jianyao dmslamafeaturebasedslamsystemforrigiddynamicscenes |