tuantdang / v3d-slam

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

V3D-SLAM: Robust RGB-D SLAM in Dynamic Environments with 3D Semantic Geometry Voting

Authors

  1. Tuan Dang
  2. Khang Nguyen
  3. Manfred Huber

All authors are with Learning and Adaptive Robotics Laboratory, Department of Computer Science and Engineering, University of Texas at Arlington, Arlington, TX 76013, USA.

Abstract

Simultaneous localization and mapping (SLAM) in highly dynamic environments is challenging due to the correlation complexity between moving objects and the camera pose. Many methods have been proposed to deal with this problem; however, the moving properties of dynamic objects with a moving camera remain unclear. Therefore, to improve SLAM's performance, minimizing disruptive events of moving objects with a physical understanding of 3D shapes and dynamics of objects is needed. In this paper, we propose a robust method, V3D-SLAM, to remove moving objects via two lightweight re-evaluation stages, including identifying potentially moving and static objects using a spatial-reasoned Hough voting mechanism and refining static objects by detecting dynamic noise caused by intra-object motions using Chamfer distances as similarity measurements. Through our experiment on the TUM RGB-D benchmark on dynamic sequences with ground-truth camera trajectories, the results show that our methods outperform most other recent state-of-the-art SLAM methods.

Full demo is available at YouTube.


Baxter is mounted with Intel RealSense D435i RGB-D

Pipeline Overview

Overview of V3D-SLAM: improving the robustness of RGB-D SLAM in dynamic indoor environments, including instance segmentation coupled with RGB-based feature extraction, sensor noises and segmentation outlier rejection, and spatial-reasoned Hough voting mechanism for dynamic 3D objects, resulting in camera trajectory estimation.


Evaluation on TUM RGB-D Dataset


Comparisons of ATE between RGB-D SLAM techniques


Comparisons of Translational Drift in RPE between RGB-D SLAM techniques


Comparisons of Rotational Drift in RPE between RGB-D SLAM techniques

Citing

About


Languages

Language:C++ 95.4%Language:Python 3.9%Language:CMake 0.4%Language:Shell 0.2%