Abstract
Most of the current visual Simultaneous Localization and Mapping (SLAM) algorithms are designed based on the assumption of a static environment, and their robustness and accuracy in the dynamic environment do not behave well. The reason is that moving objects in the scene will cause the mismatch of features in the pose estimation process, which further affects its positioning and mapping accuracy. In the meantime, the three-dimensional semantic map plays a key role in mobile robot navigation, path planning, and other tasks. In this paper, we present OFM-SLAM: Optical Flow combining MASK-RCNN SLAM, a novel visual SLAM for semantic mapping in dynamic indoor environments. Firstly, we use the Mask-RCNN network to detect potential moving objects which can generate masks of dynamic objects. Secondly, an optical flow method is adopted to detect dynamic feature points. Then, we combine the optical flow method and the MASK-RCNN for full dynamic points’ culling, and the SLAM system is able to track without these dynamic points. Finally, the semantic labels obtained from MASK-RCNN are mapped to the point cloud for generating a three-dimensional semantic map that only contains the static parts of the scenes and their semantic information. We evaluate our system in public TUM datasets. The results of our experiments demonstrate that our system is more effective in dynamic scenarios, and the OFM-SLAM can estimate the camera pose more accurately and acquire a more precise localization in the high dynamic environment.
Highlights
Simultaneous Localization and Mapping (SLAM) enables the mobile robot to estimate the current position and posture through the sensor and the corresponding motion estimation algorithm without any prior environmental information and establish a three-dimensional map of the environment
Based on the complex and dynamic indoor dynamic environment, this paper explores the methods of constructing the semantic map in the dynamic environment, combining the visual SLAM system based on the RGB-D camera and the deep learning method. e methods proposed in this paper have a certain value in semantic map construction in the dynamic environment and can help robots achieve more intelligent navigation tasks
We proposed the OFM-SLAM system which is based on the state-of-the-art Oriented Fast and Rotated Binary Robust Independent Elementary Features (BRIEF) (ORB)-SLAM2
Summary
Simultaneous Localization and Mapping (SLAM) enables the mobile robot to estimate the current position and posture through the sensor and the corresponding motion estimation algorithm without any prior environmental information and establish a three-dimensional map of the environment. RGB-D cameras have become one of the important sensors for the mobile robot assembly due to its cost effectiveness, application occasions, and availability of rich scene information. Dynamic objects in the environment, such as walking people, opening and closing doors, or any other change in the environment, will bring unpredictable abnormal observations to the system, reduce the positioning accuracy of mobile robots, cause dynamic interfering objects to become part of the environment map, and even cause the SLAM system to fail completely. Erefore, the existing algorithms are not well applicable to dynamic environments, and the accuracy and robustness of SLAM systems in dynamic environments need to be improved. Intelligent mobile robots need to have a higher level of understanding of the scene to perform complex tasks, such as the semantic information of surrounding objects and their location information. e
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
