Fusing Panoptic Segmentation and Geometry Information for Robust Visual SLAM in Dynamic Environments

Zhu，Hu1,2; Yao，Chen1,2; Zhu，Zheng1,2; Liu，Zhengtao1,2; Jia，Zhenzhong1,2

10.1109/CASE49997.2022.9926478

2022

2161-8070

2161-8089

978-1-6654-9043-6

IEEE International Conference on Automation Science and Engineering

2022-August

1648-1653

20-24 Aug. 2022

Mexico City, Mexico

Mobile robots need reliable maps for autonomous operation. Traditional SLAM systems, which are mainly developed for static scenes, often fail in dynamic environments with moving objects present in the scene. Learning based dynamic SLAM systems often suffer from insufficient or inaccurate identification of feature points. This paper proposes a novel real-time RGB-D SLAM system, which is targeted for dynamic environments, can further enhance feature detection and dynamic removal. This is done by fusing panoptic segmentation and geometry information. The system includes four components: dynamic segmentation that reduces the impact of moving objects, pose estimation with dynamic object removal, panoptic mapping, and loop closing. The pose estimation uses coarse-to-fine dynamic/static classification to further reduce the impact of unknown moving objects. Extensive evaluations demonstrate that our system can achieve robust performance in complex dynamic environments.

Geometry Visualization Simultaneous localization and mapping Computer aided software engineering Automation Feature detection Pose estimation

First ; Corresponding

English

EI

20224613111558

2-s2.0-85141676499

Scopus

1.Southern University of Science and Technology (SUSTech),Shenzhen Key Laboratory of Biomimetic Robotics and Intelligent Systems,China
2.Guangdong Prov. Key Laboratory of Human-Augmentation and Rehabilitation Robotics in Universities,Department of Mechanical and Energy Engineering,Shenzhen,518055,China

Zhu，Hu,Yao，Chen,Zhu，Zheng,et al. Fusing Panoptic Segmentation and Geometry Information for Robust Visual SLAM in Dynamic Environments[C],2022:1648-1653.