Abstract
Visual Simultaneous Localization and Mapping (SLAM) system is mainly used in real-time localization and mapping tasks of robots in various complex environments, while traditional monocular vision algorithms are struggling to cope with weak texture and dynamic scenes. To solve these problems, this work presents an object detection and clustering assisted SLAM algorithm (OC-SLAM), which adopts a faster object detection algorithm to add semantic information to the image and conducts geometrical constraint on the dynamic keypoints in the prediction box to optimize the camera pose. It also uses RGB-D camera to perform dense point cloud reconstruction with the dynamic objects rejected, and facilitates European clustering of dense point clouds to jointly eliminate dynamic features combining with object detection algorithm. Experiments in the TUM dataset indicate that OC-SLAM enhances the localization accuracy of the SLAM system in the dynamic environments compared with original algorithm and it has shown impressive performance in the localizition and can build a more precise dense point cloud map in dynamic scenes.
Highlights
The indoor mobile robot is a robot system composed of multi-sensor fusion perception, autonomous decision making, mission planning, and control, etc
Bescos et al present a dynamic Simultaneous Localization and Mapping (SLAM) system based on ORBSLAM2 (Mur-Artal and Tardós, 2017) with Mask-RCNN semantic segmentation (Bescos et al, 2018), which contains monocular, binocular, and RGB-D inputs, and the extracted dynamic ORB features are rejected by invoking the Mask-RCNN model, but this system is mainly time-consuming in the semantic segmentation algorithm and cannot achieve real-time pose estimation
Inspired by recent researches based on the semantic algorithm, we investigate the problem of real-time localization and dense map construction for the indoor mobile robots and propose a novel RGB-D SLAM framework which leverages a faster object detection method to obtain semantic information from RGB image and perform a dense map constuction with dynamic objects rejected
Summary
The indoor mobile robot is a robot system composed of multi-sensor fusion perception, autonomous decision making, mission planning, and control, etc. Doherty et al build an IMU sensor based, semantic segmentation SLAM system which introduces data association into the SLAM system optimization process and performs land marker optimization, camera pose estimation and semantic information association simultaneously (Doherty et al, 2020) Their approaches are fail to meet the demand for real-time operation and the single semantic segmentation algorithm does not guarantee the robustness of the SLAM system in the complex operating environment of the robot. Inspired by recent researches based on the semantic algorithm, we investigate the problem of real-time localization and dense map construction for the indoor mobile robots and propose a novel RGB-D SLAM framework which leverages a faster object detection method to obtain semantic information from RGB image and perform a dense map constuction with dynamic objects rejected.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.