Abstract:Pose estimation is a crucial technique to achieve robot localization using natural features for an RGB-D vision navigation system. However, it still suffers from the impoverishment of robustness and accuracy in complex workspaces. A robust real-time pose estimation approach using hybrid point and twin line reprojection is proposed to allow accurate localization for a vision- based robot under various challenging scenarios, including low texture, low structure, occlusion, and viewpoint and illumination changes. First, an improved adaptive threshold (IAT)-based oriented fast and rotated brief (ORB) feature detection method is developed to extract sufficient IAT-ORB point features that are matched by a multistage K-nearest neighbor (KNN)-ratio-test criterion (RTC)-cosine similarity (CS)-random sample consensus (RANSAC) (KRCR) refinement feature matching strategy to generate reliable point matches. Then, a robust EDLines line feature matching method is developed by integrating dynamic matching (DM) and median absolute deviation (MAD) to obtain fine line matches. Furthermore, the hybrid point reprojection errors are built from the fusion of 3-D-2-D and 3-D-3-D point pairs, and the twin line reprojection errors are constructed by integrating virtual right-eye lines generated by depth measurement, which both are incorporated into a novel unified errors' optimization model to achieve optimal motion estimation by minimizing the reprojection errors of points and lines concurrently. Finally, the extensive experiments verify the robustness, accuracy, and real-time performance of the proposed approach on the public datasets and our real indoor scenarios. The quantitative results demonstrate that our proposed approach gains more accurate and stable localization than other state-ofthe-art methods under challenging environments.

Robust Depth-Aided RGBD-Inertial Odometry for Indoor Localization

A Framework for Multi-Session RGBD SLAM in Low Dynamic Workspace Environment.

Robust Indoor Localization and Map Matching Algorithm with Visual SLAM

R2DIO: A Robust and Real-Time Depth-Inertial Odometry Leveraging Multimodal Constraints for Challenging Environments

Feature Regions Segmentation Based RGB-D Visual Odometry in Dynamic Environment

R 2 DIO: A Robust and Real-Time Depth-Inertial Odometry Leveraging Multi-Modal Constraints for Challenging Environments

A Robust RGB-D SLAM System With Points and Lines for Low Texture Indoor Environments

Robust Depth-Aided Visual-Inertial-Wheel Odometry for Mobile Robots

Robust Keyframe-based Dense SLAM with an RGB-D Camera.

RGB‐D SLAM with Moving Object Tracking in Dynamic Environments

Real-Time Optimization-Based Dense Mapping System of RGBD-Inertial Odometry

Spatiotemporal Visual Odometry Using Ground Plane in Dynamic Indoor Environment

RGB-D Inertial Odometry for Indoor Robot via Keyframe-based Nonlinear Optimization

Visual Odometry via Hybrid Motion Estimation Method

Robust Pose Estimation via Hybrid Point and Twin Line Reprojection for RGB-D Vision Navigation

VID-SLAM: Robust Pose Estimation with RGBD-Inertial Input for Indoor Robotic Localization

Visual Odometry and 3D Mapping in Indoor Environments

Fast visual odometry and mapping from RGB-D data

RGB-D Inertial Odometry for a Resource-Restricted Robot in Dynamic Environments

Robust localization for planar moving robot in changing environment: A perspective on density of correspondence and depth

Enhancing Visual Odometry with Estimated Scene Depth: Leveraging RGB-D Data with Deep Learning